Xpersona
T

Token Optimizer for OpenClaw & Hermes

CLAWHUB · Unverified

Find potential token waste in OpenClaw or Hermes Agent. Use when asked to reduce AI costs, review model choices, check large instructions or audit recurring tasks. Explain what matters and suggest safe next steps without changing settings.

62
Rank
84
Safety
14k
Downloads

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 13.8K downloads reported by the source. Last updated 11/16/2026.

Avoid when

  • Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Install and run

Setup complexity: low.

clawhub skill install s172w5rqenpwn4qb4g5n563hx9845890:openclaw-token-optimizer
  1. Install using `clawhub skill install s172w5rqenpwn4qb4g5n563hx9845890:openclaw-token-optimizer` in an isolated environment before connecting it to live workloads.
  2. No published capability contract is available yet, so validate auth and request/response behavior manually.
  3. Review the upstream CLAWHUB listing at https://clawhub.ai/asif2bd/openclaw-token-optimizer before using production credentials.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-asif2bd-openclaw-token-optimizer-ptimizer-2/snapshot"

Run-check

$0.02 USD

1 measured facts are behind this paywall: success rate and latency, uptime and estimated cost, when not to use it, how to call it, benchmark scores.

Agents pay $0.02 in USDC. A card payment is $0.50, the smallest a card allows.

Documentation

CLAWHUB

154,974 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: openclaw-token-optimizer
description: "Find potential token waste in OpenClaw or Hermes Agent. Use when asked to reduce AI costs, review model choices, check large instructions or audit recurring tasks. Explain what matters and suggest safe next steps without changing settings."
version: 4.1.1
author: Asif2BD
homepage: https://missiondeck.ai
source: https://github.com/Asif2BD/OpenClaw-Token-Optimizer
license: MIT
openclaw: ">=2026.9.4"
metadata: {"openclaw":{"emoji":"💰","requires":{"bins":["python3"]}}}
---

# Make your AI agent leaner—not less capable

Ask your agent:

> “Check where I might be wasting tokens and tell me what to improve.”

Or try:

- “Review my scheduled tasks for unnecessary AI usage.”
- “Which instruction files are making my context heavy?”
- “Help me choose a suitable model for routine work.”

Works with **OpenClaw and Hermes Agent**. Your agent handles the technical steps below;
you do not need to copy Python commands or prepare exports when the required evidence
is already available in its environment.

## What you get

A short, prioritized explanation of what was checked, what deserves attention and what
to do next. Nothing silently changes your models, instructions or schedules. Potential
savings are opportunities to test—not a promise or a measured reduction in your bill.

Built by [MissionDeck.ai](https://missiondeck.ai).

## Instructions for the agent

### 1. Start with what you already know

When the user requests an audit, proceed with relevant read-only checks; do not ask them
to restate the runtime, agent ID or active profile if trusted session context already
provides it. Say briefly that you will inspect and recommend, without changing settings.

Use this decision order:

- An explicit user-selected runtime/profile takes priority.
- Otherwise use the hosting session's known runtime and active agent/profile.
- Installed binaries alone are not proof of which profile is active. If the host is unknown
  or multiple profiles are plausible, ask one focused question rather than auditing the
  wrong system. Continue any independent checks whose scope is already clear.

Resolve bundled script paths relative to this SKILL.md directory. Use an available Python
3.10+ interpreter. No automatic installs, recursive home-directory searches or credential
searches. Never print raw configuration or environment values.

### 2. Run the matching read-only audit

**In OpenClaw:** use the current agent ID from trusted runtime context, not the example
name or the CLI's default. Follow `references/QUICKSTART.md` for the command. Include
`--config` only when the active config path is already established and the audit needs it.
If the CLI is unavailable, use exports already supplied; otherwise report the missing
capability instead of asking the user to run a long list of commands.

**In Hermes:** use the active profile path already supplied by the host/user. A known
HERMES_HOME path is profile evidence; do not dump the environmen

README.md

# Token Optimizer for OpenClaw & Hermes

[![Version](https://img.shields.io/badge/version-4.1.1-brightgreen.svg)](CHANGELOG.md)
[![MissionDeck](https://img.shields.io/badge/MissionDeck-ai-blueviolet)](https://missiondeck.ai)
[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE.txt)

Built by [MissionDeck.ai](https://missiondeck.ai) · [GitHub](https://github.com/Asif2BD/OpenClaw-Token-Optimizer) · [ClawHub](https://clawhub.ai/asif2bd/skills/openclaw-token-optimizer)

**Find where your AI agent may be wasting tokens—and get a clear, safe improvement plan.**

Large instructions, poorly matched models and repetitive AI tasks can all add overhead.
Token Optimizer helps your OpenClaw or Hermes agent investigate those opportunities and
explain what is worth changing. It does not silently change your setup or promise savings
it has not measured.

## Start by asking your agent

Once the skill is installed, say:

> “Use Token Optimizer to check my setup and suggest the three most useful improvements.”

You can also ask:

- “Are my scheduled tasks using AI when a simple script would do?”
- “Which instruction files are taking up the most context?”
- “Review my model choices for everyday tasks.”
- “Check my Hermes profile for unnecessary recurring AI work.”

**You do not need to begin with terminal commands.** A tool-enabled agent uses its known
host runtime, current agent or active profile to perform read-only checks. It asks only
when information is missing or ambiguous. The skill cannot add terminal access to an
agent that does not have it.

## What you will get

- **A clear verdict:** what was inspected and what deserves attention.
- **Prioritized suggestions:** a few useful next steps, with evidence and caveats.
- **No surprise changes:** your models, schedules and instruction files stay untouched.
- **Honest limits:** missing spending data stays unknown, not a misleading zero.

For example, it may flag an agent task running every few minutes for review, or identify
large instruction files. Neither proves waste by itself; your agent explains the trade-off
before recommending a change. Actual savings require a comparable before/after measurement.

## One skill, two platforms

**OpenClaw:** inspects the native model catalog/status and scheduled jobs. Optional known
config files add deeper checks. Recognizes agent tasks, commands, heartbeats and system events.

**Hermes Agent:** reads the active profile's configuration and scheduled jobs. Traces model
choices including saved creation-time defaults, skips paused jobs, and recognizes no-agent
scripts. This is file-based auditing—not native model discovery or authentication testing.

**Both:** can inspect selected instruction sizes, duplicate content and supplied usage
figures. Model recommendations need a user policy and sufficient availability evidence.

The agent selects the runtime from trusted context. If both platforms are installed and
no active runtime is known, it asks rather than 

_meta.json

{
  "ownerId": "kn78dz5s9fev1b0vez1kxz1fj580ckv2",
  "slug": "openclaw-token-optimizer",
  "version": "4.1.1",
  "publishedAt": 1789834857525
}

references/HERMES.md

# Hermes Agent compatibility

The read-only adapter was verified against NousResearch/hermes-agent commit
`d7b836ab1c0cddaafc109ed24c9a83b6191cdc88`.
Source: https://github.com/NousResearch/hermes-agent/tree/d7b836ab1c0cddaafc109ed24c9a83b6191cdc88

Maintainer integration: `python3 tools/test_hermes_compat.py --source /path/to/pinned/checkout`.
This creates an isolated temporary profile using the actual upstream cron.jobs parse/save
functions, audits the resulting files, and verifies every profile file remains unchanged.
No scheduler, job script, provider authentication or paid inference is run.

Inputs: explicit profile config.yaml and cron/jobs.json, or JSON/YAML config plus JSON jobs export.
Missing selected files fail instead of reporting an empty successful audit. The adapter does
not import Hermes, search profiles, read .env or resolve profile overlays. Ambient HERMES_MODEL
is ignored; pass its known non-secret value with --hermes-model-env if needed.

Model order follows upstream cron scheduler: job.model, cron.model, model_snapshot,
explicit environment, global model.default (or model string). Provider evidence follows
job.provider, cron.model_provider, provider_snapshot, global model.provider. Actual auth,
aliases, runtime overrides and fallbacks remain unverified.

The skill format uses SKILL.md. Preserve its scripts and references when installing.
OpenClaw native commands are never invoked for a Hermes audit. JSON needs standard-library
Python 3.10+; YAML additionally needs PyYAML, loaded with safe_load. Nothing is auto-installed.

No native catalog/auth/usage ingestion, automatic configuration changes or guaranteed savings.
Reports omit raw prompt/script/config content; selected identifiers can still be sensitive.

references/PROVIDERS.md

# Provider policy

Do not infer credentials from environment variables or hard-code a cheapest model.
Use the selected agent's native catalog/status and an operator-owned ordered tier policy.
The bundled policy is an optional OpenAI example, not a universal recommendation.
Change the identifiers for the models your account exposes.

Availability, allowlist membership, capability metadata, and successful inference are
separate evidence. An available catalog row does not prove that OAuth, a particular
runtime, or the exact model route will work. v4 never runs a paid canary automatically.
Before activation, explicitly test the exact route and inspect the actual model used.

Unknown image/tool/context capability blocks a requirement that needs it. Catalogs
that omit tool metadata require a trustworthy normalized export with supportsTools=true;
do not fill this field by guessing. Native route issues block recommendations until resolved.

Tier ordering represents the user's quality/cost/latency preference. v4 does not scrape
prices, claim live pricing, infer model quality from names, or estimate model latency.
Retirements should be handled by refreshing catalog/status and reviewing provider notices.

Prompt cache lifetime and billing vary by provider, account, and model. Compare measured
cache writes, cache reads, input and output cost. Never assume that extra cache-warming
heartbeats save money. Same-provider fallbacks do not guarantee resilience to auth outages.

Sources: https://docs.openclaw.ai/releases/2026.9.4
https://developers.openai.com/codex/models

Sponsored

Ads related to Token Optimizer for OpenClaw & Hermes and adjacent AI workflows.