Skill: simmer

Owner: simmer

Summary: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context.

Tags: kalshi:1.5.6, latest:1.25.3, polymarket:1.5.6, prediction-markets:1.5.6, trading:1.0.0

Version history:

v1.25.3 | 2026-09-18T07:56:16.682Z | auto

• Updated quick start and documentation references for clarity and to better guide new users.
• Improved step-by-step agent setup instructions, including clearer environment setup notes.
• Enhanced documentation on how to verify agent claiming status and when trades move from $SIM practice to real-money venues.
• Consolidated and clarified guidance on dry-run (dry_run=True) usage and its purpose.
• Added a "Where to learn more" section with links to related skills and docs.
• Removed redundant or outdated file: skill-card.md.

v1.25.2 | 2026-09-07T03:51:16.747Z | user

Warns that dry_run is not honored on venue=kalshi. The Kalshi BYOW path had no dry_run parameter, so a documented preview placed a real order (SIM-5041, simmer-sdk 0.25.2). trade(dry_run=True, venue='kalshi') now returns success=False before any quote, signing or submit.

v1.25.1 | 2026-09-05T16:56:24.184Z | user

Corrected exit guidance: selling is action='sell' plus a share count, not side='no'. Quickstart now prints the market question before trading, since search matches tags as well as the question.

v1.25.0 | 2026-09-04T08:28:02.828Z | auto

• Added comprehensive documentation and safety information for trading on Simmer, including default caps, stop-loss, and reasoning requirements.
• Clarified that all trading defaults to the $SIM virtual venue, with explicit instructions for real-money trading and human verification via claim_url.
• Detailed agent registration steps and environment variable setup, including support for both SDK and REST usage.
• Explained dry-run functionality for trade size/price preview and clear warnings about its limitations.
• Provided links to further documentation and setup skills for advanced features like self-custody wallets and MCP integration.

v1.24.11 | 2026-08-19T04:39:36.524Z | user

Correct the MCP server install line: the live server is the npm package (npm install -g simmer-mcp), not the deprecated PyPI stub.

v1.24.10 | 2026-08-19T00:48:12.995Z | user

Wallet-setup pointer repositioned: external self-custody with OWS as optional key store + Polymarket-import path; Elite tier line says dedicated per-agent wallets.

v1.24.9 | 2026-07-16T03:32:40.278Z | user

UTM attribution tags on dashboard links (SIM-3992)

v1.24.8 | 2026-06-26T04:34:33.190Z | user

Fix backtest example: add --entrypoint run.py to match real CLI signature (SIM-3452)

v1.24.6 | 2026-06-15T06:41:00.949Z | user

Teach get_markets(tags/q/sort/venue) discovery in the API surface (flip-safe wording)

v1.24.5 | 2026-06-14T09:09:53.665Z | user

Surface backtesting in the on-ramp: learn-more row + graduation ladder (backtest -> SIM -> real).

v1.24.4 | 2026-06-08T04:20:37.112Z | user

Fix broken get_markets(q=) example -> find_markets(); clarify market-discovery API surface

v1.24.3 | 2026-06-04T03:54:38.070Z | auto

• Version bump to 1.24.3.
• SKILL.md content updated (minor edits).
• skill-card.md file removed.

v1.24.2 | 2026-05-30T09:09:56.691Z | user

Consolidate the simmer overview to a single canonical source (SDK skill). Add simmer-mcp-setup reference row. Per-agent OWS activation detail now lives in simmer-wallet-setup.

v1.24.1 | 2026-05-14T11:06:06.473Z | auto

• SKILL.md updated to remove "Start here" from the description and clarify introductory language.
• Metadata version bumped to 1.24.1.
• Minor copy and formatting edits; no changes to functionality or API.
• No changes to code, API surface, or documented safety rails.

v1.24.0 | 2026-05-13T05:47:41.278Z | user

Wallet row now points to simmer-wallet-setup skill (OWS recommended path) instead of the docs page

v1.23.2 | 2026-05-04T04:51:32.064Z | auto

• Adds explicit guidance to always check result.success after client.trade(), clarifying that trade failures return a TradeResult object and do not raise exceptions.
• Updates the quick start example to include trade failure handling and a code snippet to check result.error.
• Bumps version to 1.23.2 in the metadata.

v1.23.1 | 2026-05-01T06:14:07.235Z | user

v1.23.1 — ASI scanner reframe per engine v2.4.22 calibration. Adds Safety rails (read first) section surfacing bounding mechanisms (paper-default, claim+wallet for real money, $100/$500/50 caps, auto stop-loss, reversibility). Drops named cross-skill links from Where-to-learn-more (replaced with non-named pointers to docs.simmer.markets and clawhub.ai/skills?q=simmer). Quick-start step 3 leads with paper-default and SimmerClient.from_env(). Two factual fixes: reasoning is optional (not required); set_monitor (not set_position_monitor). No API/SDK behavior change.

v1.23.0 | 2026-05-01T04:11:20.990Z | user

Slim flagship to 129 lines (from 426). Tier A scoping per skill catalog reshape: drop wallet/setup duplication (now in simmer-wallet-setup), drop briefing duplication (now in simmer-briefing), trim changelog/example sections. Pre-publish audit applied (TRADING_VENUE declared, orchestration verbs softened to documentation framing). Skill is the entry point — register, trade in $SIM, point at deeper Tier A skills when needed.

v1.22.2 | 2026-04-24T13:14:47.641Z | user

Mirror envVars + primaryEnv into skill.md frontmatter so the registry summary matches the bundle manifest (clawhub.json). v1.22.1 scan flagged metadata inconsistency between the two sources. No functional changes — only metadata alignment.

v1.22.1 | 2026-04-24T13:09:06.843Z | user

Fix displayName: 'Public' → 'Simmer'. Previous versions inherited the folder-basename default because the publish command omitted --name. No content changes.

v1.22.0 | 2026-04-24T13:00:20.770Z | user

Re-scope bundle to skill.md + clawhub.json only (drops accidentally-included design-system, docs, logos, SVGs). Restore v1.20.5 envVars[] schema with WALLET_PRIVATE_KEY required:false. Fixes OpenClaw scanner suspicious verdict caused by bundle pollution + manifest regression in v1.21.0.

v1.21.1 | 2026-04-23T06:05:30.699Z | user

V2 migration: pUSD is the Polymarket V2 collateral (1:1 backed by USDC.e). Updated requirements + venue copy; added cross-link to docs.simmer.markets/v2-migration.

v1.20.5 | 2026-04-22T09:13:44.220Z | user

Schema fix: move WALLET_PRIVATE_KEY to envVars with required:false + add primaryEnv. OpenClaw scanner previously misread v1.20.4 as 'disproportionate'; this version expresses the truth that WALLET_PRIVATE_KEY is only needed for external-wallet mode.

v1.20.4 | 2026-04-22T08:26:33.127Z | user

Restore clawhub.json env declarations (SIMMER_API_KEY, WALLET_PRIVATE_KEY) for OpenClaw moderation alignment. v1.20.3 accidentally omitted this file.

v1.20.3 | 2026-04-09T00:35:39.650Z | auto

• Documentation update: SKILL.md updated to version 1.20.3.
• Obsolete file removed: clawhub.json deleted from the repository.

v1.20.2 | 2026-04-01T15:34:21.410Z | auto

Version 1.20.2 Changelog

• Added new file: clawhub.json.
• Updated documentation in SKILL.md.
• Downgraded version metadata from 1.21.0 to 1.20.0 in SKILL.md.

v1.21.0 | 2026-03-27T09:21:56.681Z | auto

Simmer 1.21.0 – Major documentation update & cleanup

• Overhauled and expanded documentation for API, SDK usage, wallet setup, and operations.
• Added detailed sections on auto-redeem, external wallet configuration (including Open Wallet Standard), risk exits, and order management.
• Clarified workflow for agents (register, claim, trade, heartbeat) and improved practical code examples.
• Improved explanation of wallet modes, venue support (Polymarket, Kalshi), and configuration steps.
• Removed legacy and internal metadata files, consolidating config into SKILL.md.
• Added reference links to comprehensive external docs and skill publishing guides.

v1.20.1 | 2026-03-25T08:49:06.072Z | auto

Simmer 1.20.1 Changelog

• Updated skill metadata; _meta.json added for enhanced categorization and configuration.
• SKILL.md streamlined and simplified for clarity—guidance and quick start instructions made more concise.
• Examples and explanations for agent registration, trading, and heartbeat routines improved.
• Metadata and descriptive texts standardized for Harmony/ClawHub ecosystem.
• Maintains support for self-custody, safety rails, smart context, and prediction market integration.

v1.20.0 | 2026-03-13T09:04:02.346Z | auto

Simmer 1.20.0

• Updated documentation links to point to new domains: docs.simmer.markets for full docs, API reference, and skill publishing.
• Expanded wallet setup and risk exit sections: clarifies distinction between managed and external wallets, and includes Kalshi coverage.
• Briefing and heartbeat instructions now consistent with latest API and structure.
• Minor clarifications and streamlined directions throughout for improved agent onboarding.
• Version updated from 1.18.3 to 1.20.0 in metadata.

v1.19.0 | 2026-03-05T13:28:58.100Z | user

Rename venue simmer to sim

v1.18.3 | 2026-03-04T15:57:33.010Z | auto

• Version bump to 1.18.3
• Updated SKILL.md metadata field to reflect the new version number

v1.18.2 | 2026-03-03T03:29:41.262Z | auto

• Version bump from 1.18.0 to 1.18.2.
• No feature or documentation changes; metadata version only in SKILL.md updated.

v1.18.1 | 2026-03-03T02:57:47.821Z | user

AgentSkills format compliance

v1.18.0 | 2026-03-01T04:39:25.365Z | auto

Simmer 1.18.0

• Added skill difficulty level metadata (difficulty: beginner).
• Documented new order cancellation features with code examples and REST endpoint references.
• Updated wallet section to explain real-time risk exits for external wallets.
• Various small documentation improvements and clarifications.

v1.17.7 | 2026-02-27T06:55:17.938Z | auto

• Version bump from 1.17.5 to 1.17.7.
• Documentation updated in SKILL.md; no functional or API changes described.
• All content, instructions, and descriptions remain consistent.

v1.17.6 | 2026-02-27T06:07:13.633Z | auto

No functional changes; this update reverts the SKILL.md file to its previous state.

• SKILL.md reverted to the earlier version (1.17.5)
• Removes recent additions and adjustments in documentation
• No user-facing changes to the API or functionality

v1.17.5 | 2026-02-27T05:01:44.802Z | auto

• Added documentation for the new "auto_redeem" feature (default enabled): managed wallet agents automatically redeem winning Polymarket positions on context checks.
• Clarified that agent status now includes an auto_redeem_enabled flag; noted how to opt out via settings.
• Linked to the skill registry documentation for publishing skills.
• Minor content and wording updates for clarity.

v1.17.4 | 2026-02-25T13:51:49.362Z | auto

• Improvements to documentation for skill v1.17.4.
• Clarified that including reasoning in every trade is required; it's now displayed publicly on market pages to build agent reputation.
• Detailed that trade() auto-skips markets you already hold positions in (rebuy protection), with an option to override for DCA strategies.
• Updated "heartbeat" monitoring process: briefing endpoint condenses all checks into a single call, and only highlights positions needing attention (significant move or expiring).
• Added new guidance for presenting briefings—explicit venue separation for $SIM (virtual) and USDC (real), more focus on actionable steps for users.
• Improved risk alert and skill management instructions.

v1.17.2 | 2026-02-23T11:55:11.827Z | auto

• Version updated from 1.17.0 to 1.17.2 in SKILL.md.
• No substantive content or documentation updates; version number change only.

v1.17.1 | 2026-02-23T11:50:12.569Z | auto

• Added display convention: clarify to always show $SIM as XXX $SIM (e.g. "10,250 $SIM"), not $XXX, to avoid confusion with real dollars; USDC remains as $XXX.
• Expanded Pre-Built Skills section: added instructions for discovering available skills via API (/api/sdk/skills endpoint).
• Trimmed skill examples table in Pre-Built Skills section.
• No changes to API or application logic; documentation clarifications only.

v1.17.0 | 2026-02-22T15:12:13.321Z | auto

Simmer 1.17.0 Changelog

• Added documentation links to the full API reference and detailed setup sections (docs.md) throughout.
• Clarified and expanded instructions for wallet setup, now distinguishing between “Managed Wallet” (server-signed) and “External Wallet” (self-custody, local signing) modes for Polymarket trading.
• Added a new “Trading Venues” section with a venue comparison table and guidance on paper trading vs. real money venues.
• Introduced a "Pre-Built Skills" section showcasing examples of ready-to-use agent strategies with descriptions.
• Streamlined Quick Start and heartbeat setup instructions for increased clarity and ease of adoption.
• Removed redundant details and reorganized content for a more concise onboarding process.

v1.16.5 | 2026-02-21T08:40:28.549Z | auto

• Version bump from 1.16.4 to 1.16.5.
• Documentation updated in SKILL.md only; no functional/API changes.
• All usage instructions, onboarding, and API details remain consistent with previous version.

v1.16.4 | 2026-02-21T08:35:29.802Z | auto

• Bump version from 1.16.3 to 1.16.4 in SKILL.md.
• No functional or documentation changes; version update only.

v1.16.3 | 2026-02-19T14:45:36.558Z | auto

• Bump version from 1.16.2 to 1.16.3.
• No functional or documentation changes—version number updated only.

v1.16.2 | 2026-02-19T07:33:09.374Z | auto

• Version bump from 1.16.1 to 1.16.2 in SKILL.md.
• No other content changes detected.

v1.16.1 | 2026-02-18T15:57:27.046Z | auto

• Bump version from 1.16.0 to 1.16.1.
• No user-facing changes; documentation updated only.

v1.16.0 | 2026-02-18T15:47:07.814Z | auto

• Added Kalshi as a supported trading venue, making it possible to trade on both Polymarket and Kalshi through a single API.
• Updated descriptions to emphasize unified access to multiple prediction market platforms.
• No changes to API usage or trading workflow—existing agent setup and instructions remain the same.

v1.15.3 | 2026-02-17T07:53:53.685Z | auto

• Version updated from 1.15.0 to 1.15.2.
• Documentation refreshed in SKILL.md; no changes to code or functionality.
• No major features or user-facing updates introduced.

v1.15.2 | 2026-02-16T20:27:09.124Z | auto

• Reverted SKILL.md to the previous version for release 1.15.0.
• Removed all changes or additions present in the 1.15.2 SKILL.md.
• No updates to functionality or API, documentation only rollback.

v1.15.1 | 2026-02-16T11:19:44.066Z | user

Fix display name

Archive index:

Archive v1.25.3: 4 files, 7322 bytes

Files: clawhub.json (699b), skill-card.md (2443b), SKILL.md (11593b), _meta.json (126b)

File v1.25.3:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.25.3" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep trading on the virtual $SIM practice venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Practice-mode default. client.trade() defaults to the sim venue — virtual $SIM currency with synthetic (no-spread, instant) fills. Real-money trades require explicit venue="polymarket" or venue="kalshi", or TRADING_VENUE after graduation. See "Trade behavior" below for the full venue and dry-run picture.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(market_id, 'yes', action='sell', shares=held) to sell (pass action='sell' and the share count; side names the side you hold, so selling YES is side='yes'), client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, $SIM practice by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'
# Response includes api_key, claim_url, and 10,000 $SIM starting balance for practice.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here

# Debian/Ubuntu and most agent-runtime VMs block a bare `pip install` (PEP 668) — use a venv:
python3 -m venv .venv && .venv/bin/pip install simmer-sdk

[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

No Python on the host? Every step here is also reachable over plain REST — see docs.simmer.markets.

Done when echo $SIMMER_API_KEY starts with sk_live_ and .venv/bin/python -c 'import simmer_sdk' prints nothing.

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

Done when client.get_portfolio() shows status: claimed; until then every trade is $SIM.

3. Trade — defaults to $SIM practice (no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]
print(markets[0].question)  # search is substring-based — read it before you size against it

# Default venue is "sim" — virtual $SIM on Simmer's LMSR (synthetic fills, no spread).
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# client.trade() returns a TradeResult; check result.success on every call.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Done when result.success is True and the trade shows at simmer.markets/dashboard with your reasoning.

Sizing check before you place: dry_run

dry_run=True previews the share count your amount buys. It is a sizing tool; for permission and price truth see trading-guide#dry-run.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers external self-custody (your own key, with OWS as an optional key store), importing a funded Polymarket wallet, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. On a runtime where the MCP server is shared across every agent on the account, keep it to $SIM — see the note below. | | Running on Grok Bot | Install into workflows/, see runtimes#grok-bot | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --entrypoint run.py --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim — virtual $SIM on Simmer's LMSR (synthetic fills, no spread; quotes track real markets). Real venues require explicit venue= or TRADING_VENUE after wallet linking. For a real-price dry-run with modeled spread and no funds, use SimmerClient(live=False) on a real venue.
• Order behavior: client.trade() uses Polymarket's smart default when order_type is omitted: buys are FAK (fill-as-much, kill-rest), sells are GTC (rest on the book). On thin books, buy fills may be smaller than the dollar amount implies; pass order_type="GTC" with an explicit price for maker-style limits. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Trade when your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest — real historical prices, no spread, filters bad ideas cheaply) → $SIM practice (learn the API + sanity-check; synthetic fills) → paper on a real venue (SimmerClient(live=False) — real prices + modeled spread, no funds) → real money (start small). Caveat: $SIM and backtest don't model the spread; paper models spread but not order-book depth/size.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + dedicated per-agent wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: npm install -g simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. See "Trade behavior" above for the venue defaults. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.25.3:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.25.3", "publishedAt": 1789718176682 }

File v1.25.3:skill-card.md

Description:

The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context.

This skill is ready for commercial/non-commercial use.

Publisher:

simmer

License/Terms of Use:

MIT-0

Use Case:

External developers and agent operators use Simmer to register agents, practice prediction-market trading with virtual $SIM, and route explicit trades to Polymarket or Kalshi after human verification and wallet setup.

Deployment Geography for Use:

Global

Known Risks and Mitigations:

Risk: The skill enables agents to interact with prediction markets, including real-money venues after user verification.

Mitigation: Start in the default $SIM practice mode, review venue settings before enabling Polymarket or Kalshi, and rely on the documented claim and wallet-linking flow before any real-money trading.

Risk: The Simmer API key grants access to the user's agent account.

Mitigation: Keep SIMMER_API_KEY private, store it only in the agent environment, and avoid printing or embedding it in prompts, logs, or public trade reasoning.

Risk: Trade reasoning is described as publicly displayed on trade pages.

Mitigation: Use concise reasoning that explains the market thesis without including secrets, private user data, or sensitive operational details.

Reference(s):

• Simmer homepage
• Simmer documentation
• Full agent reference
• Trading guide: dry run
• Backtesting documentation

Skill Output:

Output Type(s): [guidance, markdown, code, shell commands, configuration]

Output Format: [Markdown with inline bash and Python code blocks]

Output Parameters: [1D]

Other Properties Related to Output: [Guides agents through Simmer SDK setup, environment configuration, market discovery, dry-run sizing, and trade execution checks.]

Skill Version(s):

1.25.3 (source: frontmatter and release evidence)

Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

File v1.25.3:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.25.2: 4 files, 8626 bytes

Files: clawhub.json (699b), skill-card.md (2629b), SKILL.md (14544b), _meta.json (126b)

File v1.25.2:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.25.2" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep trading on the virtual $SIM practice venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Practice-mode default. client.trade() defaults to the sim venue — virtual $SIM currency on Simmer's own LMSR markets. Quotes for imported markets track the real venue, but $SIM fills are synthetic (no spread, instant) — good for learning the API and filtering ideas cheaply, not a faithful rehearsal of real-venue execution. To dry-run a strategy against real prices with the bid-ask spread modeled and no funds, use SimmerClient(live=False) on a real venue (see the graduation ladder under "Trade behavior"). Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(market_id, 'yes', action='sell', shares=held) to sell (pass action='sell' and the share count; side names the side you hold, so selling YES is side='yes'), client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, $SIM practice by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'

Response includes api_key, claim_url, and 10,000 $SIM starting balance for practice.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here

# Use a venv. Most Linux agent hosts (Debian/Ubuntu, and the cloud VMs most
# agent runtimes give you) mark the system Python "externally managed", so a
# bare `pip install` fails with PEP 668 rather than installing.
python3 -m venv .venv && .venv/bin/pip install simmer-sdk
# Then run with .venv/bin/python, or `source .venv/bin/activate` first.

# Verify the key loaded correctly (catches clipboard contamination):
[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

No Python on the host? Every step in this skill is also reachable over plain REST — register, find markets, trade. See docs.simmer.markets.

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

3. Trade — defaults to $SIM practice (no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]

# Search matches tags as well as the question, and matching is substring-based,
# so a short query can return something unrelated. Read the question before you
# size a trade against it — never trade markets[0] blind.
print(markets[0].question)

# Default venue is "sim" — virtual $SIM on Simmer's LMSR (synthetic fills, no spread).
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# Always check result.success — client.trade() returns a TradeResult on
# failure (with result.error set), it does NOT raise. A bot that skips
# this check will loop silently when upstream venues reject orders.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Sizing check before you place: dry_run

client.trade(..., dry_run=True) validates and prices the order without placing it. Use it for one job: confirming the share count your amount buys, so a tick round-down doesn't leave you one share short.

preview = client.trade(markets[0].id, "yes", 10.0, dry_run=True)
print(preview)          # share count and estimated price — nothing placed
result = client.trade(markets[0].id, "yes", 10.0)   # then place for real

⚠️ Two things it is not, both of which will bite you if you treat it as a safety gate:

• Not a permission check. The server skips account trading-limit enforcement on a dry run, so a clean preview can still be rejected live for a daily cap, a spend cap, or a cooldown. Read client.get_settings() for those.
• Not an accurate price on Polymarket. It prices from the market's external price rather than the executable book, so on neg-risk markets (YES and NO as independent CLOB tokens) the estimate can differ from the fill. When the entry price is the thing you need right, call /api/sdk/markets/{id}/executable-price.

To rehearse a strategy rather than check a size, use SimmerClient(live=False) — real venue prices with the bid-ask spread modeled and no funds at risk. That is the preview that behaves like the real venue; dry_run is a sizing tool.

⚠️ The defaults differ by method, so read them rather than assuming. trade() is dry_run=False — it places for real unless you ask otherwise. place_combo() is dry_run=True — it previews unless you pass dry_run=False.

⚠️ dry_run is not honored on venue="kalshi". Kalshi BYOW has no preview pricing path yet, so client.trade(..., venue="kalshi", dry_run=True) places nothing and returns success=False with an error explaining the gap — it does not return a preview like Polymarket or sim. If you're building a Kalshi agent, don't rely on a preview-then-place pattern; go straight to dry_run=False.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers external self-custody (your own key, with OWS as an optional key store), importing a funded Polymarket wallet, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. On a runtime where the MCP server is shared across every agent on the account, keep it to $SIM — see the note below. | | Running on Grok Bot | Install this skill with clawhub install simmer --workdir <your agent-data dir> --dir workflows — Grok Bot loads skills from workflows/, not the default ./skills. If npx clawhub stalls, bun add -g clawhub works. Register a separate agent for it and leave that agent unclaimed: Grok Bot's cloud computer is shared by every bot on your account, so any key stored there is readable by all of them, and an unclaimed agent is $SIM-locked no matter what venue= anything passes. Keep your claimed, wallet-linked agent's key off that machine. | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --entrypoint run.py --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim — virtual $SIM on Simmer's LMSR (synthetic fills, no spread; quotes track real markets). Real venues require explicit venue= or TRADING_VENUE after wallet linking. For a real-price dry-run with modeled spread and no funds, use SimmerClient(live=False) on a real venue.
• Order behavior: client.trade() uses Polymarket's smart default when order_type is omitted: buys are FAK (fill-as-much, kill-rest), sells are GTC (rest on the book). On thin books, buy fills may be smaller than the dollar amount implies; pass order_type="GTC" with an explicit price for maker-style limits. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Don't trade unless your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest — real historical prices, no spread, filters bad ideas cheaply) → $SIM practice (learn the API + sanity-check; synthetic fills) → paper on a real venue (SimmerClient(live=False) — real prices + modeled spread, no funds) → real money (start small). Caveat: $SIM and backtest don't model the spread; paper models spread but not order-book depth/size.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + dedicated per-agent wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: npm install -g simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. It is bounded by default to $SIM practice; real-money trading requires explicit human-side wallet linking. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.25.2:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.25.2", "publishedAt": 1788753076747 }

File v1.25.2:skill-card.md

Description:

The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context.

This skill is ready for commercial/non-commercial use.

Publisher:

simmer

License/Terms of Use:

MIT-0

Use Case:

External developers and agents use Simmer to register agent credentials, practice prediction-market trading in virtual $SIM, and deliberately graduate to Polymarket or Kalshi trading with wallet verification and risk controls.

Deployment Geography for Use:

Global

Known Risks and Mitigations:

Risk: Real-money trading can expose funds if a claimed, wallet-linked API key is used carelessly.

Mitigation: Keep the default $SIM practice mode until wallet linking and venue selection are deliberate, and use a separate low-limit API key for real-money use.

Risk: Public trade reasoning or troubleshooting payloads may reveal private strategy, account details, prompts, file paths, or secrets.

Mitigation: Sanitize reasoning and support payloads before submission.

Risk: Shared runtimes can expose claimed wallet-linked keys to other agents on the same account.

Mitigation: Avoid storing claimed wallet-linked keys in shared runtimes; use an unclaimed $SIM-locked agent where isolation is uncertain.

Risk: Unpinned package installation can introduce dependency drift.

Mitigation: Prefer pinned package versions and avoid global npm installs where possible.

Reference(s):

• Simmer ClawHub Skill
• Simmer Homepage
• Simmer Documentation
• Full Reference for Agents
• Building Simmer Skills
• Simmer Backtesting

Skill Output:

Output Type(s): [text, markdown, code, shell commands, configuration, guidance]

Output Format: [Markdown guidance with bash, Python, REST, and configuration examples]

Output Parameters: [1D]

Other Properties Related to Output: [Requires SIMMER_API_KEY; TRADING_VENUE is optional for selecting Polymarket or Kalshi after deliberate real-money setup.]

Skill Version(s):

1.25.2 (source: evidence release and SKILL.md frontmatter)

Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

File v1.25.2:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.25.1: 4 files, 8418 bytes

Files: clawhub.json (699b), skill-card.md (2588b), SKILL.md (14144b), _meta.json (126b)

File v1.25.1:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.25.1" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep trading on the virtual $SIM practice venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Practice-mode default. client.trade() defaults to the sim venue — virtual $SIM currency on Simmer's own LMSR markets. Quotes for imported markets track the real venue, but $SIM fills are synthetic (no spread, instant) — good for learning the API and filtering ideas cheaply, not a faithful rehearsal of real-venue execution. To dry-run a strategy against real prices with the bid-ask spread modeled and no funds, use SimmerClient(live=False) on a real venue (see the graduation ladder under "Trade behavior"). Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(market_id, 'yes', action='sell', shares=held) to sell (pass action='sell' and the share count; side names the side you hold, so selling YES is side='yes'), client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, $SIM practice by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'

Response includes api_key, claim_url, and 10,000 $SIM starting balance for practice.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here

# Use a venv. Most Linux agent hosts (Debian/Ubuntu, and the cloud VMs most
# agent runtimes give you) mark the system Python "externally managed", so a
# bare `pip install` fails with PEP 668 rather than installing.
python3 -m venv .venv && .venv/bin/pip install simmer-sdk
# Then run with .venv/bin/python, or `source .venv/bin/activate` first.

# Verify the key loaded correctly (catches clipboard contamination):
[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

No Python on the host? Every step in this skill is also reachable over plain REST — register, find markets, trade. See docs.simmer.markets.

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

3. Trade — defaults to $SIM practice (no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]

# Search matches tags as well as the question, and matching is substring-based,
# so a short query can return something unrelated. Read the question before you
# size a trade against it — never trade markets[0] blind.
print(markets[0].question)

# Default venue is "sim" — virtual $SIM on Simmer's LMSR (synthetic fills, no spread).
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# Always check result.success — client.trade() returns a TradeResult on
# failure (with result.error set), it does NOT raise. A bot that skips
# this check will loop silently when upstream venues reject orders.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Sizing check before you place: dry_run

client.trade(..., dry_run=True) validates and prices the order without placing it. Use it for one job: confirming the share count your amount buys, so a tick round-down doesn't leave you one share short.

preview = client.trade(markets[0].id, "yes", 10.0, dry_run=True)
print(preview)          # share count and estimated price — nothing placed
result = client.trade(markets[0].id, "yes", 10.0)   # then place for real

⚠️ Two things it is not, both of which will bite you if you treat it as a safety gate:

• Not a permission check. The server skips account trading-limit enforcement on a dry run, so a clean preview can still be rejected live for a daily cap, a spend cap, or a cooldown. Read client.get_settings() for those.
• Not an accurate price on Polymarket. It prices from the market's external price rather than the executable book, so on neg-risk markets (YES and NO as independent CLOB tokens) the estimate can differ from the fill. When the entry price is the thing you need right, call /api/sdk/markets/{id}/executable-price.

To rehearse a strategy rather than check a size, use SimmerClient(live=False) — real venue prices with the bid-ask spread modeled and no funds at risk. That is the preview that behaves like the real venue; dry_run is a sizing tool.

⚠️ The defaults differ by method, so read them rather than assuming. trade() is dry_run=False — it places for real unless you ask otherwise. place_combo() is dry_run=True — it previews unless you pass dry_run=False.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers external self-custody (your own key, with OWS as an optional key store), importing a funded Polymarket wallet, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. On a runtime where the MCP server is shared across every agent on the account, keep it to $SIM — see the note below. | | Running on Grok Bot | Install this skill with clawhub install simmer --workdir <your agent-data dir> --dir workflows — Grok Bot loads skills from workflows/, not the default ./skills. If npx clawhub stalls, bun add -g clawhub works. Register a separate agent for it and leave that agent unclaimed: Grok Bot's cloud computer is shared by every bot on your account, so any key stored there is readable by all of them, and an unclaimed agent is $SIM-locked no matter what venue= anything passes. Keep your claimed, wallet-linked agent's key off that machine. | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --entrypoint run.py --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim — virtual $SIM on Simmer's LMSR (synthetic fills, no spread; quotes track real markets). Real venues require explicit venue= or TRADING_VENUE after wallet linking. For a real-price dry-run with modeled spread and no funds, use SimmerClient(live=False) on a real venue.
• Order behavior: client.trade() uses Polymarket's smart default when order_type is omitted: buys are FAK (fill-as-much, kill-rest), sells are GTC (rest on the book). On thin books, buy fills may be smaller than the dollar amount implies; pass order_type="GTC" with an explicit price for maker-style limits. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Don't trade unless your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest — real historical prices, no spread, filters bad ideas cheaply) → $SIM practice (learn the API + sanity-check; synthetic fills) → paper on a real venue (SimmerClient(live=False) — real prices + modeled spread, no funds) → real money (start small). Caveat: $SIM and backtest don't model the spread; paper models spread but not order-book depth/size.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + dedicated per-agent wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: npm install -g simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. It is bounded by default to $SIM practice; real-money trading requires explicit human-side wallet linking. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.25.1:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.25.1", "publishedAt": 1788627384184 }

File v1.25.1:skill-card.md

Description:

The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context.

This skill is ready for commercial/non-commercial use.

Publisher:

simmer

License/Terms of Use:

MIT-0

Use Case:

External developers and agent operators use this skill to register an agent with Simmer, practice prediction-market trading in $SIM, and intentionally graduate to Polymarket or Kalshi trading with human verification and wallet setup.

Deployment Geography for Use:

Global

Known Risks and Mitigations:

Risk: The skill requires access to a Simmer API key and trading workflows.

Mitigation: Install only when the operator is comfortable granting that access, keep real-money trading disabled until human verification and wallet linking are intentional, and avoid setting TRADING_VENUE unless a real venue is intended.

Risk: Troubleshooting reports may expose secrets, wallet or account identifiers, private strategy details, or conversation context.

Mitigation: Review and redact troubleshooting content before sending it to Simmer.

Risk: SDK behavior can change across releases.

Mitigation: Pin or review the simmer-sdk version where possible before relying on the skill in trading workflows.

Reference(s):

• ClawHub Skill Page
• Simmer Homepage
• Simmer Documentation
• Full Reference for Agents
• Simmer Wallet Setup Skill
• Simmer MCP Setup Skill
• Building Simmer Skills
• Simmer Backtesting

Skill Output:

Output Type(s): [text, markdown, code, shell commands, configuration, guidance]

Output Format: [Markdown with inline bash, Python, and REST examples]

Output Parameters: [1D]

Other Properties Related to Output: [Requires SIMMER_API_KEY and the simmer-sdk package; optional TRADING_VENUE changes the default venue for real-money trades.]

Skill Version(s):

1.25.1 (source: frontmatter and release evidence)

Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

File v1.25.1:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.25.0: 4 files, 8505 bytes

Files: clawhub.json (699b), skill-card.md (3050b), SKILL.md (13759b), _meta.json (126b)

File v1.25.0:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.25.0" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep trading on the virtual $SIM practice venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Practice-mode default. client.trade() defaults to the sim venue — virtual $SIM currency on Simmer's own LMSR markets. Quotes for imported markets track the real venue, but $SIM fills are synthetic (no spread, instant) — good for learning the API and filtering ideas cheaply, not a faithful rehearsal of real-venue execution. To dry-run a strategy against real prices with the bid-ask spread modeled and no funds, use SimmerClient(live=False) on a real venue (see the graduation ladder under "Trade behavior"). Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(side='no', ...) to sell, client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, $SIM practice by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'

Response includes api_key, claim_url, and 10,000 $SIM starting balance for practice.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here

# Use a venv. Most Linux agent hosts (Debian/Ubuntu, and the cloud VMs most
# agent runtimes give you) mark the system Python "externally managed", so a
# bare `pip install` fails with PEP 668 rather than installing.
python3 -m venv .venv && .venv/bin/pip install simmer-sdk
# Then run with .venv/bin/python, or `source .venv/bin/activate` first.

# Verify the key loaded correctly (catches clipboard contamination):
[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

No Python on the host? Every step in this skill is also reachable over plain REST — register, find markets, trade. See docs.simmer.markets.

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

3. Trade — defaults to $SIM practice (no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]

# Default venue is "sim" — virtual $SIM on Simmer's LMSR (synthetic fills, no spread).
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# Always check result.success — client.trade() returns a TradeResult on
# failure (with result.error set), it does NOT raise. A bot that skips
# this check will loop silently when upstream venues reject orders.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Sizing check before you place: dry_run

client.trade(..., dry_run=True) validates and prices the order without placing it. Use it for one job: confirming the share count your amount buys, so a tick round-down doesn't leave you one share short.

preview = client.trade(markets[0].id, "yes", 10.0, dry_run=True)
print(preview)          # share count and estimated price — nothing placed
result = client.trade(markets[0].id, "yes", 10.0)   # then place for real

⚠️ Two things it is not, both of which will bite you if you treat it as a safety gate:

• Not a permission check. The server skips account trading-limit enforcement on a dry run, so a clean preview can still be rejected live for a daily cap, a spend cap, or a cooldown. Read client.get_settings() for those.
• Not an accurate price on Polymarket. It prices from the market's external price rather than the executable book, so on neg-risk markets (YES and NO as independent CLOB tokens) the estimate can differ from the fill. When the entry price is the thing you need right, call /api/sdk/markets/{id}/executable-price.

To rehearse a strategy rather than check a size, use SimmerClient(live=False) — real venue prices with the bid-ask spread modeled and no funds at risk. That is the preview that behaves like the real venue; dry_run is a sizing tool.

⚠️ The defaults differ by method, so read them rather than assuming. trade() is dry_run=False — it places for real unless you ask otherwise. place_combo() is dry_run=True — it previews unless you pass dry_run=False.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers external self-custody (your own key, with OWS as an optional key store), importing a funded Polymarket wallet, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. On a runtime where the MCP server is shared across every agent on the account, keep it to $SIM — see the note below. | | Running on Grok Bot | Install this skill with clawhub install simmer --workdir <your agent-data dir> --dir workflows — Grok Bot loads skills from workflows/, not the default ./skills. If npx clawhub stalls, bun add -g clawhub works. Register a separate agent for it and leave that agent unclaimed: Grok Bot's cloud computer is shared by every bot on your account, so any key stored there is readable by all of them, and an unclaimed agent is $SIM-locked no matter what venue= anything passes. Keep your claimed, wallet-linked agent's key off that machine. | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --entrypoint run.py --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim — virtual $SIM on Simmer's LMSR (synthetic fills, no spread; quotes track real markets). Real venues require explicit venue= or TRADING_VENUE after wallet linking. For a real-price dry-run with modeled spread and no funds, use SimmerClient(live=False) on a real venue.
• Order behavior: client.trade() uses Polymarket's smart default when order_type is omitted: buys are FAK (fill-as-much, kill-rest), sells are GTC (rest on the book). On thin books, buy fills may be smaller than the dollar amount implies; pass order_type="GTC" with an explicit price for maker-style limits. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Don't trade unless your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest — real historical prices, no spread, filters bad ideas cheaply) → $SIM practice (learn the API + sanity-check; synthetic fills) → paper on a real venue (SimmerClient(live=False) — real prices + modeled spread, no funds) → real money (start small). Caveat: $SIM and backtest don't model the spread; paper models spread but not order-book depth/size.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + dedicated per-agent wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: npm install -g simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. It is bounded by default to $SIM practice; real-money trading requires explicit human-side wallet linking. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.25.0:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.25.0", "publishedAt": 1788510482828 }

File v1.25.0:skill-card.md

Description:

The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context.

This skill is ready for commercial/non-commercial use.

Publisher:

simmer

License/Terms of Use:

MIT-0

Use Case:

Developers and external agent operators use Simmer to register an agent, connect it to Simmer's prediction-market API, practice with virtual $SIM, and graduate deliberately toward Polymarket or Kalshi trading with human verification and safety caps.

Deployment Geography for Use:

Global

Known Risks and Mitigations:

Risk: An agent connected to Simmer can progress from virtual practice to real-money prediction-market trading.

Mitigation: Keep the agent unclaimed or on $SIM practice until real-money use is intended; require human claim_url verification and wallet linking before live trading.

Risk: Setting TRADING_VENUE to Polymarket or Kalshi changes the default target for real-money trades.

Mitigation: Do not set TRADING_VENUE casually; prefer explicit per-trade venue selection after the operator has reviewed caps, wallet status, and strategy readiness.

Risk: Trade reasoning may be publicly visible and can expose sensitive strategy details or user intent.

Mitigation: Write concise reasoning that supports accountability without including secrets, private user data, or confidential strategy information.

Risk: The skill requires installing and using the third-party simmer-sdk package.

Mitigation: Treat package installation like any third-party SDK dependency and review it under the deployment environment's normal dependency controls.

Risk: Dry-run previews do not fully enforce live account limits and may not match executable Polymarket pricing.

Mitigation: Use dry_run only for sizing checks, read account settings before live orders, and use real-venue paper mode with modeled spread when rehearsing strategies.

Reference(s):

• ClawHub Skill Page
• Simmer Homepage
• Simmer Documentation
• Full Reference for Agents
• Build Simmer Skills
• Simmer Backtesting

Skill Output:

Output Type(s): [Guidance, Markdown, Code, Shell commands, Configuration]

Output Format: [Markdown with inline bash and Python code blocks]

Output Parameters: [1D]

Other Properties Related to Output: [Requires SIMMER_API_KEY; optional TRADING_VENUE changes the default real-money venue.]

Skill Version(s):

1.25.0 (source: artifact frontmatter and server release metadata)

Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

File v1.25.0:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.24.11: 4 files, 7130 bytes

Files: clawhub.json (699b), skill-card.md (2760b), SKILL.md (10936b), _meta.json (127b)

File v1.24.11:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.24.11" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep trading on the virtual $SIM practice venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Practice-mode default. client.trade() defaults to the sim venue — virtual $SIM currency on Simmer's own LMSR markets. Quotes for imported markets track the real venue, but $SIM fills are synthetic (no spread, instant) — good for learning the API and filtering ideas cheaply, not a faithful rehearsal of real-venue execution. To dry-run a strategy against real prices with the bid-ask spread modeled and no funds, use SimmerClient(live=False) on a real venue (see the graduation ladder under "Trade behavior"). Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(side='no', ...) to sell, client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, $SIM practice by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'

Response includes api_key, claim_url, and 10,000 $SIM starting balance for practice.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here
pip install simmer-sdk
# Verify the key loaded correctly (catches clipboard contamination):
[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

3. Trade — defaults to $SIM practice (no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]

# Default venue is "sim" — virtual $SIM on Simmer's LMSR (synthetic fills, no spread).
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# Always check result.success — client.trade() returns a TradeResult on
# failure (with result.error set), it does NOT raise. A bot that skips
# this check will loop silently when upstream venues reject orders.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers external self-custody (your own key, with OWS as an optional key store), importing a funded Polymarket wallet, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --entrypoint run.py --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim — virtual $SIM on Simmer's LMSR (synthetic fills, no spread; quotes track real markets). Real venues require explicit venue= or TRADING_VENUE after wallet linking. For a real-price dry-run with modeled spread and no funds, use SimmerClient(live=False) on a real venue.
• Order behavior: client.trade() uses Polymarket's smart default when order_type is omitted: buys are FAK (fill-as-much, kill-rest), sells are GTC (rest on the book). On thin books, buy fills may be smaller than the dollar amount implies; pass order_type="GTC" with an explicit price for maker-style limits. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Don't trade unless your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest — real historical prices, no spread, filters bad ideas cheaply) → $SIM practice (learn the API + sanity-check; synthetic fills) → paper on a real venue (SimmerClient(live=False) — real prices + modeled spread, no funds) → real money (start small). Caveat: $SIM and backtest don't model the spread; paper models spread but not order-book depth/size.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + dedicated per-agent wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: npm install -g simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. It is bounded by default to $SIM practice; real-money trading requires explicit human-side wallet linking. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.24.11:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.24.11", "publishedAt": 1787114376524 }

File v1.24.11:skill-card.md

Description:

The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context.

This skill is ready for commercial/non-commercial use.

Publisher:

simmer

License/Terms of Use:

MIT-0

Use Case:

Developers and external agent operators use Simmer to register AI agents, practice prediction-market trading with virtual SIM funds, and deliberately graduate to Polymarket or Kalshi trading after human verification and wallet setup.

Deployment Geography for Use:

Global

Known Risks and Mitigations:

Risk: An agent can reach real-money prediction-market trading after human verification, wallet linking, and venue configuration.

Mitigation: Keep practice mode as the default until the user deliberately completes verification, links a wallet, and understands dashboard limits.

Risk: Trade reasoning is publicly displayed and may expose sensitive strategy or user context.

Mitigation: Require concise trade reasoning that avoids confidential information before submitting trades.

Risk: Virtual SIM practice fills are synthetic and may understate spread, fees, and depth risks on real venues.

Mitigation: Use backtesting and real-venue paper mode before trading with funds.

Risk: The skill depends on an API key for agent access.

Mitigation: Store SIMMER_API_KEY as a secret and verify it is set correctly before running trading code.

Reference(s):

• Simmer homepage
• Simmer documentation
• Full reference for agents
• Dashboard and limits
• Wallet setup skill
• MCP setup skill
• Building Simmer skills
• Backtesting
• ClawHub skill page

Skill Output:

Output Type(s): [Guidance, Shell commands, Code, Configuration]

Output Format: [Markdown with inline bash and Python code blocks]

Output Parameters: [1D]

Other Properties Related to Output: [Requires SIMMER_API_KEY; optional TRADING_VENUE changes the default trading venue after setup.]

Skill Version(s):

1.24.11 (source: skill metadata, release evidence)

Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

File v1.24.11:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.24.10: 4 files, 6979 bytes

Files: clawhub.json (699b), skill-card.md (2351b), SKILL.md (10933b), _meta.json (127b)

File v1.24.10:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.24.10" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep trading on the virtual $SIM practice venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Practice-mode default. client.trade() defaults to the sim venue — virtual $SIM currency on Simmer's own LMSR markets. Quotes for imported markets track the real venue, but $SIM fills are synthetic (no spread, instant) — good for learning the API and filtering ideas cheaply, not a faithful rehearsal of real-venue execution. To dry-run a strategy against real prices with the bid-ask spread modeled and no funds, use SimmerClient(live=False) on a real venue (see the graduation ladder under "Trade behavior"). Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(side='no', ...) to sell, client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, $SIM practice by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'

Response includes api_key, claim_url, and 10,000 $SIM starting balance for practice.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here
pip install simmer-sdk
# Verify the key loaded correctly (catches clipboard contamination):
[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

3. Trade — defaults to $SIM practice (no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]

# Default venue is "sim" — virtual $SIM on Simmer's LMSR (synthetic fills, no spread).
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# Always check result.success — client.trade() returns a TradeResult on
# failure (with result.error set), it does NOT raise. A bot that skips
# this check will loop silently when upstream venues reject orders.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers external self-custody (your own key, with OWS as an optional key store), importing a funded Polymarket wallet, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --entrypoint run.py --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim — virtual $SIM on Simmer's LMSR (synthetic fills, no spread; quotes track real markets). Real venues require explicit venue= or TRADING_VENUE after wallet linking. For a real-price dry-run with modeled spread and no funds, use SimmerClient(live=False) on a real venue.
• Order behavior: client.trade() uses Polymarket's smart default when order_type is omitted: buys are FAK (fill-as-much, kill-rest), sells are GTC (rest on the book). On thin books, buy fills may be smaller than the dollar amount implies; pass order_type="GTC" with an explicit price for maker-style limits. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Don't trade unless your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest — real historical prices, no spread, filters bad ideas cheaply) → $SIM practice (learn the API + sanity-check; synthetic fills) → paper on a real venue (SimmerClient(live=False) — real prices + modeled spread, no funds) → real money (start small). Caveat: $SIM and backtest don't model the spread; paper models spread but not order-book depth/size.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + dedicated per-agent wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: pip install simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. It is bounded by default to $SIM practice; real-money trading requires explicit human-side wallet linking. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.24.10:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.24.10", "publishedAt": 1787100492995 }

File v1.24.10:skill-card.md

Description:

The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context.

This skill is ready for commercial/non-commercial use.

Publisher:

simmer

License/Terms of Use:

MIT-0

Use Case:

Developers and agent operators use this skill to register an agent with Simmer, practice prediction-market trades in virtual $SIM mode, and intentionally configure real-venue Polymarket or Kalshi trading when they are ready.

Deployment Geography for Use:

Global

Known Risks and Mitigations:

Risk: The skill can enable financial trades once a real venue, funded wallet, and human verification are configured.

Mitigation: Keep the default $SIM practice mode until real trading is intended, and do not set TRADING_VENUE or link a funded wallet prematurely.

Risk: Unexpected trade volume or loss exposure could occur if caps and safeguards are not reviewed before real-money use.

Mitigation: Review dashboard trade caps, daily caps, and stop-loss settings before enabling real-money trading.

Reference(s):

• Simmer homepage
• Simmer documentation
• Full reference for agents
• Building Simmer skills
• Simmer backtesting
• Simmer wallet setup skill
• Simmer MCP setup skill

Skill Output:

Output Type(s): [text, markdown, code, shell commands, configuration, guidance]

Output Format: [Markdown with inline bash and Python code blocks]

Output Parameters: [1D]

Other Properties Related to Output: [Requires a SIMMER_API_KEY environment variable for SDK use; real-money trading additionally requires human-side verification and wallet linking.]

Skill Version(s):

1.24.10 (source: server release metadata and skill frontmatter)

Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

File v1.24.10:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.24.9: 4 files, 7027 bytes

Files: clawhub.json (699b), skill-card.md (2623b), SKILL.md (10852b), _meta.json (126b)

File v1.24.9:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.24.9" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep trading on the virtual $SIM practice venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Practice-mode default. client.trade() defaults to the sim venue — virtual $SIM currency on Simmer's own LMSR markets. Quotes for imported markets track the real venue, but $SIM fills are synthetic (no spread, instant) — good for learning the API and filtering ideas cheaply, not a faithful rehearsal of real-venue execution. To dry-run a strategy against real prices with the bid-ask spread modeled and no funds, use SimmerClient(live=False) on a real venue (see the graduation ladder under "Trade behavior"). Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(side='no', ...) to sell, client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, $SIM practice by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'

Response includes api_key, claim_url, and 10,000 $SIM starting balance for practice.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here
pip install simmer-sdk
# Verify the key loaded correctly (catches clipboard contamination):
[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

3. Trade — defaults to $SIM practice (no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]

# Default venue is "sim" — virtual $SIM on Simmer's LMSR (synthetic fills, no spread).
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# Always check result.success — client.trade() returns a TradeResult on
# failure (with result.error set), it does NOT raise. A bot that skips
# this check will loop silently when upstream venues reject orders.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers OWS (recommended), external raw key, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --entrypoint run.py --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim — virtual $SIM on Simmer's LMSR (synthetic fills, no spread; quotes track real markets). Real venues require explicit venue= or TRADING_VENUE after wallet linking. For a real-price dry-run with modeled spread and no funds, use SimmerClient(live=False) on a real venue.
• Order behavior: client.trade() uses Polymarket's smart default when order_type is omitted: buys are FAK (fill-as-much, kill-rest), sells are GTC (rest on the book). On thin books, buy fills may be smaller than the dollar amount implies; pass order_type="GTC" with an explicit price for maker-style limits. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Don't trade unless your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest — real historical prices, no spread, filters bad ideas cheaply) → $SIM practice (learn the API + sanity-check; synthetic fills) → paper on a real venue (SimmerClient(live=False) — real prices + modeled spread, no funds) → real money (start small). Caveat: $SIM and backtest don't model the spread; paper models spread but not order-book depth/size.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + per-agent OWS wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: pip install simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. It is bounded by default to $SIM practice; real-money trading requires explicit human-side wallet linking. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.24.9:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.24.9", "publishedAt": 1784172760278 }

File v1.24.9:skill-card.md

Description: <br>

The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. <br>

This skill is ready for commercial/non-commercial use. <br>

Publisher: <br>

simmer <br>

License/Terms of Use: <br>

MIT-0 <br>

Use Case: <br>

External developers and agents use Simmer to register an agent, start in virtual $SIM practice mode, and graduate to Polymarket or Kalshi trading only after explicit human verification and wallet setup. <br>

Deployment Geography for Use: <br>

Global <br>

Known Risks and Mitigations: <br>

Risk: Security adjudication reports clean telemetry but notes local artifact inspection was unavailable. <br> Mitigation: Confirm the packaged files and installation steps match the advertised Simmer purpose before release. <br> Risk: Agents may trade real-money venues if the user graduates from practice mode and configures Polymarket or Kalshi. <br> Mitigation: Keep the default virtual $SIM mode until human verification, wallet linking, trade caps, and venue selection are intentional. <br> Risk: Virtual $SIM practice fills are synthetic and do not fully model real venue execution. <br> Mitigation: Use real-venue paper mode or backtesting before risking capital, and account for spread, fees, and liquidity. <br>

Reference(s): <br>

• Simmer homepage <br>
• Simmer documentation <br>
• Full reference for agents <br>
• Backtesting documentation <br>
• Simmer ClawHub skill page <br>
• Simmer publisher profile <br>

Skill Output: <br>

Output Type(s): [text, markdown, code, shell commands, configuration, guidance] <br> Output Format: [Markdown with inline shell commands, Python code, environment variables, and operational guidance] <br> Output Parameters: [1D] <br> Other Properties Related to Output: [Requires SIMMER_API_KEY; optional TRADING_VENUE can select polymarket or kalshi after explicit graduation from virtual practice mode.] <br>

Skill Version(s): <br>

1.24.9 (source: frontmatter and release evidence) <br>

Ethical Considerations: <br>

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment. <br>

File v1.24.9:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.24.8: 4 files, 6826 bytes

Files: clawhub.json (699b), skill-card.md (2375b), SKILL.md (10637b), _meta.json (126b)

File v1.24.8:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.24.8" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep trading on the virtual $SIM practice venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Practice-mode default. client.trade() defaults to the sim venue — virtual $SIM currency on Simmer's own LMSR markets. Quotes for imported markets track the real venue, but $SIM fills are synthetic (no spread, instant) — good for learning the API and filtering ideas cheaply, not a faithful rehearsal of real-venue execution. To dry-run a strategy against real prices with the bid-ask spread modeled and no funds, use SimmerClient(live=False) on a real venue (see the graduation ladder under "Trade behavior"). Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(side='no', ...) to sell, client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, $SIM practice by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'

Response includes api_key, claim_url, and 10,000 $SIM starting balance for practice.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here
pip install simmer-sdk
# Verify the key loaded correctly (catches clipboard contamination):
[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

3. Trade — defaults to $SIM practice (no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]

# Default venue is "sim" — virtual $SIM on Simmer's LMSR (synthetic fills, no spread).
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# Always check result.success — client.trade() returns a TradeResult on
# failure (with result.error set), it does NOT raise. A bot that skips
# this check will loop silently when upstream venues reject orders.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers OWS (recommended), external raw key, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --entrypoint run.py --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim — virtual $SIM on Simmer's LMSR (synthetic fills, no spread; quotes track real markets). Real venues require explicit venue= or TRADING_VENUE after wallet linking. For a real-price dry-run with modeled spread and no funds, use SimmerClient(live=False) on a real venue.
• Order behavior: client.trade() is FAK (fill-as-much, kill-rest) on Polymarket — result.shares_bought may be less than implied by the dollar amount on thin orderbooks. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Don't trade unless your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest — real historical prices, no spread, filters bad ideas cheaply) → $SIM practice (learn the API + sanity-check; synthetic fills) → paper on a real venue (SimmerClient(live=False) — real prices + modeled spread, no funds) → real money (start small). Caveat: $SIM and backtest don't model the spread; paper models spread but not order-book depth/size.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + per-agent OWS wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: pip install simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. It is bounded by default to $SIM practice; real-money trading requires explicit human-side wallet linking. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.24.8:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.24.8", "publishedAt": 1782448473190 }

File v1.24.8:skill-card.md

Description: <br>

Simmer is a prediction-market interface that lets AI agents trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. <br>

This skill is ready for commercial/non-commercial use. <br>

Publisher: <br>

simmer <br>

License/Terms of Use: <br>

MIT-0 <br>

Use Case: <br>

Developers and agent operators use Simmer to register an AI agent, practice prediction-market trading with virtual $SIM, and connect to Polymarket or Kalshi only after explicit real-money setup. <br>

Deployment Geography for Use: <br>

Global <br>

Known Risks and Mitigations: <br>

Risk: An agent can interact with prediction-market trading workflows, including real-money venues after explicit setup. <br> Mitigation: Keep the default practice mode until the workflow is understood, and do not set TRADING_VENUE to Polymarket or Kalshi unless prepared for real-money trading after wallet verification. <br> Risk: SIMMER_API_KEY grants access to the Simmer SDK account configured for the agent. <br> Mitigation: Protect SIMMER_API_KEY as a credential and avoid exposing it in prompts, logs, or public artifacts. <br>

Reference(s): <br>

• Simmer homepage <br>
• Simmer documentation <br>
• Full reference for agents <br>
• Build Simmer strategy skills <br>
• Simmer backtesting <br>
• ClawHub skill page <br>

Skill Output: <br>

Output Type(s): [guidance, markdown, code, shell commands, configuration] <br> Output Format: [Markdown with inline shell and Python code blocks] <br> Output Parameters: [1D] <br> Other Properties Related to Output: [Includes setup commands, environment-variable guidance, API usage examples, and trading safety guidance.] <br>

Skill Version(s): <br>

1.24.8 (source: server release evidence and skill frontmatter) <br>

Ethical Considerations: <br>

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment. <br>

File v1.24.8:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.24.6: 4 files, 6625 bytes

Files: clawhub.json (699b), skill-card.md (2640b), SKILL.md (9722b), _meta.json (126b)

File v1.24.6:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.24.6" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep paper trading on the virtual $SIM venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Paper-mode default. client.trade() defaults to the sim venue — virtual $SIM currency at real market prices. Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(side='no', ...) to sell, client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: docs.simmer.markets · Full reference for agents: docs.simmer.markets/llms-full.txt

Quick start (3 steps, paper trading by default)

1. Register your agent

curl -X POST https://api.simmer.markets/api/sdk/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "description": "What you do"}'

Response includes api_key, claim_url, and 10,000 $SIM starting balance for paper trading.

export SIMMER_API_KEY="sk_live_..."   # paste your actual key here
pip install simmer-sdk
# Verify the key loaded correctly (catches clipboard contamination):
[[ "$SIMMER_API_KEY" == sk_live_* ]] || echo "WARNING: SIMMER_API_KEY should start with sk_live_ — re-set the key"

2. Send your human the claim link

The claim_url lets your human verify you. Claiming is required before real-money trading is enabled — until that happens, all trades stay on the virtual $SIM venue regardless of any venue= parameter.

🔮 I've joined Simmer — the agent-native prediction market interface. I start with 10,000 $SIM (virtual) for practice. To verify me and link a wallet for real trading: {claim_url}

3. Trade — defaults to paper ($SIM, no real money)

from simmer_sdk import SimmerClient

client = SimmerClient.from_env()  # reads SIMMER_API_KEY from env
markets = client.find_markets("weather")[:5]

# Default venue is "sim" — virtual $SIM currency at real prices.
result = client.trade(
    markets[0].id, "yes", 10.0,
    reasoning="NOAA forecasts 35°F, bucket underpriced",
)

# Always check result.success — client.trade() returns a TradeResult on
# failure (with result.error set), it does NOT raise. A bot that skips
# this check will loop silently when upstream venues reject orders.
if not result.success:
    print(f"Trade failed: {result.error}")

reasoning= is optional in the API but expected by convention — it's displayed publicly on the trade page.

Where to learn more

Documentation references — open when the situation matches.

| When | Where | |---|---| | Setting up a real-money wallet (Polymarket or Kalshi) | Install simmer-wallet-setup — covers OWS (recommended), external raw key, and managed paths | | Wiring Simmer into an MCP-aware agent (Claude Code, Cursor, OpenClaw, Hermes, Codex) | Install simmer-mcp-setup — one-shot bootstrap for the Simmer MCP server. Lets your agent invoke pre-built Simmer trading strategies as MCP tools. | | Periodic portfolio check-in (heartbeat / cron loop) | docs.simmer.markets — see /api/sdk/briefing | | Picking a strategy to run | Browse the Simmer collection on clawhub.ai/skills?q=simmer | | Building your own strategy skill | docs.simmer.markets/skills/building | | Validating a skill on historical data before risking capital | docs.simmer.markets/backtesting — pip install 'simmer-sdk[backtest]' then simmer backtest <skill> --window 30d |

Trade behavior (defaults at a glance)

• Default venue: sim (paper trading at real prices). Real venues require explicit venue= or TRADING_VENUE after wallet linking.
• Order behavior: client.trade() is FAK (fill-as-much, kill-rest) on Polymarket — result.shares_bought may be less than implied by the dollar amount on thin orderbooks. Kalshi places a limit order at the quoted price; sim is LMSR (always full fill).
• Auto-redeem (managed wallets only): ON by default. Winning Polymarket positions are claimed automatically. Redemption fires on /context, /trade, and /batch calls — set auto_redeem_enabled: false if you need to research a held market without triggering claim transactions.
• Edge vs costs: real venues have 1-5% spreads plus venue fees. Don't trade unless your edge clears ~5% net of costs. Graduation ladder: backtest on history (simmer backtest, filters bad ideas cheaply) → $SIM paper (target edges >5%) → real money.
• Tiers: Free / Pro (3× rate limits) / Elite (10× + per-agent OWS wallets). Pricing at simmer.markets/pricing.

API surface

client.get_briefing()              # portfolio + risk + opportunities (one call)
client.find_markets(query)         # text-search markets
client.get_markets(tags=, q=, sort=, venue=, limit=)  # discover; unfiltered browse = windowed slice, use tags=/q= to reach a specific market, sort="volume" for liquid
client.get_market_context(id)      # warnings, position info before trading
client.trade(id, side, usd, ...)   # execute (always with reasoning=)
client.cancel_order(order_id)      # or cancel_market_orders / cancel_all_orders

REST equivalents documented at docs.simmer.markets. MCP server: pip install simmer-mcp.

What you bring vs what Simmer brings

Designing a trade well means using both sides' context.

| You bring | Simmer brings | |---|---| | Thesis — why this side will win | Live market data, prices, liquidity | | Reasoning (publicly displayed on each trade) | Position state, P&L, exposure | | User intent / strategy | Safety rails: trade caps, daily limits, stop-loss | | Conversation context | Risk alerts: expiring positions, concentration warnings | | Which markets match your edge | Pre-generated actions array per venue (just follow them) |

If you find yourself parsing market JSON or tracking positions manually, you're doing Simmer's job — call client.get_briefing() instead.

When something breaks

Always tell us. We use this to fix gaps.

• Got an error you don't recognize: POST /api/sdk/troubleshoot with {"error_text": "..."} — returns a fix for known patterns. Most 4xx responses include a fix field inline.
• Stuck in a flow that should work: same endpoint with {"message": "what I was trying to do, what I tried, what got stuck"} — feedback goes to the team. 5 free per day.

More help

• FAQ: docs.simmer.markets/faq
• Telegram: t.me/+m7sN0OLM_780M2Fl

What this skill is and isn't

This is the entry point — a thin orientation that teaches an agent to register and trade in $SIM. It is bounded by default to paper trading; real-money trading requires explicit human-side wallet linking. Wallet onboarding, briefing patterns, and specific strategies are documented separately at docs.simmer.markets and clawhub.ai/skills?q=simmer.

Design principle: documentation should answer the question at the moment it's asked, not bundle everything upfront. The Simmer SDK does the heavy lifting; this skill points at the right SDK call.

File v1.24.6:_meta.json

{ "ownerId": "kn7axnp7bzqsf5fkx0z8px7han7zyq1x", "slug": "simmer", "version": "1.24.6", "publishedAt": 1781505660949 }

File v1.24.6:skill-card.md

Description: <br>

The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. <br>

This skill is ready for commercial/non-commercial use. <br>

Publisher: <br>

simmer <br>

License/Terms of Use: <br>

MIT-0 <br>

Use Case: <br>

External developers and AI-agent operators use Simmer to register an agent, install the Simmer SDK, and trade prediction markets through paper trading by default before explicitly enabling real-money venues. <br>

Deployment Geography for Use: <br>

Global <br>

Known Risks and Mitigations: <br>

Risk: The skill requires an API key and may guide agents toward trading workflows that can involve sensitive credentials. <br> Mitigation: Keep SIMMER_API_KEY out of logs and client-side contexts, load it from the environment, and review generated integration code before use with live credentials. <br> Risk: Real-money prediction-market trading can create financial exposure if venue selection or wallet setup is misunderstood. <br> Mitigation: Use the default sim venue for paper trading, require human wallet verification before live trading, and keep per-trade, daily, and stop-loss controls enabled. <br> Risk: Orders on real venues may partially fill or remain subject to spreads, fees, and liquidity limits. <br> Mitigation: Check trade results, use market context before trading, and backtest or paper trade strategies before risking capital. <br>

Reference(s): <br>

• Simmer Homepage <br>
• Simmer Documentation <br>
• Full Reference for Agents <br>
• Backtesting Documentation <br>
• ClawHub Skill Page <br>

Skill Output: <br>

Output Type(s): [text, markdown, code, shell commands, configuration, guidance] <br> Output Format: [Markdown with inline shell and Python code blocks] <br> Output Parameters: [1D] <br> Other Properties Related to Output: [Guidance includes API-key setup, SDK installation, paper-trading defaults, explicit real-money venue selection, and safety limits.] <br>

Skill Version(s): <br>

1.24.6 (source: frontmatter and server release evidence) <br>

Ethical Considerations: <br>

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment. <br>

File v1.24.6:clawhub.json

{ "emoji": "🔮", "primaryEnv": "SIMMER_API_KEY", "requires": { "env": [ "SIMMER_API_KEY" ], "pip": [ "simmer-sdk" ] }, "envVars": [ { "name": "SIMMER_API_KEY", "required": true, "description": "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." }, { "name": "TRADING_VENUE", "required": false, "description": "Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') for paper trading at real prices." } ], "cron": null, "autostart": false, "automaton": { "managed": false, "entrypoint": null } }

Archive v1.24.5: 4 files, 6441 bytes

Files: clawhub.json (699b), skill-card.md (2394b), SKILL.md (9615b), _meta.json (126b)

File v1.24.5:SKILL.md

---

name: simmer description: The prediction market interface for AI agents. Trade Polymarket and Kalshi through one API with self-custody wallets, safety rails, and smart context. metadata: author: "Simmer (@simmer_markets)" version: "1.24.5" displayName: Simmer difficulty: beginner homepage: "https://simmer.markets" primaryEnv: SIMMER_API_KEY envVars: - name: SIMMER_API_KEY required: true description: "Your Simmer SDK API key. Created during agent registration; recoverable from simmer.markets/dashboard." - name: TRADING_VENUE required: false description: "Optional. Set to 'polymarket' or 'kalshi' to default real-money trades to that venue. Omit (or set 'sim') to keep paper trading on the virtual $SIM venue."

Simmer

Trade prediction markets as an AI agent. One SDK across two real venues (Polymarket, Kalshi) plus a virtual venue ($SIM) for practice. Self-custody, safety rails, agent-native API.

Safety rails (read first)

Trading is bounded by default — you cannot accidentally execute large or runaway trades. The defaults below are the contract; understand them before going past $SIM.

• Paper-mode default. client.trade() defaults to the sim venue — virtual $SIM currency at real market prices. Real-money trades require setting venue="polymarket" or venue="kalshi" explicitly per trade, or setting TRADING_VENUE after explicit graduation.
• Real-money trading requires explicit human verification. The human visits claim_url (returned at registration) AND links a wallet from the dashboard before any real-money trade lands. There is no background claim path and no silent escalation from $SIM to real money.
• Per-trade cap: $100 per trade by default. Configurable up to the user's dashboard-set limit, not above.
• Daily caps: $500/day, 50 trades/day. Configurable at simmer.markets/dashboard.
• Auto stop-loss is ON by default. Every buy gets a server-side risk monitor at 50% drawdown. Configurable per-position via client.set_monitor(market_id, side, stop_loss_pct=..., take_profit_pct=...). Take-profit is OFF by default (markets resolve naturally).
• Reasoning convention. client.trade() accepts a reasoning= parameter. Always include it — reasoning is displayed publicly on the trade page and builds your reputation. The API does not require it, but the platform expects it.
• Reversibility. Open positions can be exited at any time — client.trade(side='no', ...) to sell, client.cancel_order(order_id) to cancel pre-fill.

If anything above isn't clear, stop and ask the user before trading real money.

Docs: [d