What makes Abaddon notable?

956 downloads. Trust evidence available

How should Abaddon be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

agentCLAWHUBUnverified

Abaddon

Red team security mode for OpenClaw. Runs an adversarial audit on demand or nightly — checks exposed ports, credential leaks, file permissions, suspicious pr...

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

956

Updated

Oct 12, 2026

Version

1.0.0

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 956 downloads reported by the source. Last updated 10/12/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 12, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 12, 2026
Adoption signal: 956 downloadsadoption · observed Oct 12, 2026
Latest release: 1.0.0release · observed Feb 20, 2026
Handshake status: UNKNOWNsecurity

What changed

Newest first. Only changes that point at a source are listed.

Release 1.0.0release
Initial release. Red team / adversarial security mode for OpenClaw. Nightly + on-demand. macOS. Letter grade output.
Observed Feb 20, 2026 · source

Install and run

Setup complexity: low.

clawhub skill install s175f7knh9wa8h414k24cvcf05885t6q:abaddon

Install using `clawhub skill install s175f7knh9wa8h414k24cvcf05885t6q:abaddon` in an isolated environment before connecting it to live workloads.
No published capability contract is available yet, so validate auth and request/response behavior manually.
Review the upstream CLAWHUB listing at https://clawhub.ai/enochosbot-bot/abaddon before using production credentials.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-enochosbot-bot-abaddon/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

10,485 characters of source documentation, loaded on request.

Extracted files

4 files captured from the source.

SKILL.md

---
name: abaddon
description: Red team security mode for OpenClaw. Runs an adversarial audit on demand or nightly — checks exposed ports, credential leaks, file permissions, suspicious processes, and OpenClaw config posture. Assigns a letter grade. Built for macOS deployments.
---

# Abaddon ⚔️

Most security audits are defensive — they check what you've locked down. Abaddon runs the other direction. It thinks like an attacker. It looks for what an adversary would find, not just what you remember to check.

On demand or every night at 3:45 AM. Letter grade every time.

## What It Checks

**Network & Exposure**
- Listening ports — anything bound to 0.0.0.0 is flagged
- Gateway binding — should be loopback only
- SSH Remote Login state
- Active tunnels (ngrok, cloudflared, unexpected remote access)
- Firewall and stealth mode

**System Integrity**
- SIP, FileVault, Gatekeeper
- macOS version + pending updates
- XProtect / MRT definitions age

**OpenClaw Configuration**
- Exec security mode (full / allowlist / deny)
- Gateway auth enabled?
- Unexpected cron entries
- Unexpected plugins

**File Permissions**
- SOUL.md + AGENTS.md: root-owned, 444
- MEMORY.md, USER.md, AGENT_PROMPT.md, openclaw.json, cron/jobs.json, LaunchAgent plists: 600
- Flags anything 644 or wider on sensitive paths
- Plaintext key scan across workspace

**API Key Handling**
- Keys in Keychain or flat files?
- Keys leaking through env vars?
- Secrets in git history?
- Hardcoded tokens in .zshrc?

**Agent Behavior**
- Memory injection scan (prompt injection attempts in memory files)
- Sub-agent scope check
- Unexpected agent permissions

**Dependencies**
- Homebrew outdated (flags openclaw, ollama, node)
- npm global outdated

## Scoring

| Grade | Criteria |
|-------|----------|
| A | 0 CRITICAL, 0 HIGH |
| B | 0 CRITICAL, 1–2 HIGH |
| C | 1 CRITICAL or 3+ HIGH |
| D | 2+ CRITICAL |
| F | Active compromise indicators |

## Installation

### Step 1 — Copy the Abaddon prompt into your agent

If you have Gideon (the OpenClaw observer agent), append the red team section:

```bash
cat skills/abaddon/templates/abaddon-prompt.md >> ~/.openclaw/workspace/agents/observer/AGENT_PROMPT.md
```

If you don't have Gideon, use the standalone agent prompt:

```bash
cp skills/abaddon/templates/abaddon-prompt.md ~/.openclaw/workspace/agents/abaddon/AGENT_PROMPT.md
```

### Step 2 — Add the nightly cron

```bash
bash skills/abaddon/setup/cron-seed.sh
```

This adds a 3:45 AM CST cron job to `~/.openclaw/cron/jobs.json`. Delivers to Telegram Security topic if configured.

### Step 3 — Lock the agent prompt

```bash
chmod 600 ~/.openclaw/workspace/agents/observer/AGENT_PROMPT.md
```

Your detection playbook should never be world-readable.

## Usage

**Manual trigger** — say any of:
- "run red team"
- "run Abaddon"
- "run full assessment"
- "Abaddon report"

**Nightly** — fires automatically at 3:45 AM CST after the standard defensive audit (3:30 AM).

## Output

Every run produces two things:

1. **

_meta.json

{
  "ownerId": "kn78z941f26sn68a4s3kcetssh81gk44",
  "slug": "abaddon",
  "version": "1.0.0",
  "publishedAt": 1771560859066
}

SPEC.md

# Abaddon — Red Team Mode (merged into Gideon)
**Date:** 2026-02-19 | **Status:** Merged into agents/observer/AGENT_PROMPT.md
**Updated:** 2026-02-19 (removed random-delay trigger, added fixed nightly cron)

---

## What It Is

Abaddon is not a separate agent. It is Gideon's adversarial mode — an offensive red team scan that runs alongside the standard nightly defensive audit.

Named after the angel of the abyss. Assumes breach. Checks what an attacker would exploit.

---

## Two Modes

| Mode | Schedule | Posture | Output |
|------|----------|---------|--------|
| Gideon nightly (defensive) | 3:30 AM CST daily | Passive monitor, internal audit | memory/audits/YYYY-MM-DD.md |
| Abaddon nightly (offensive) | 3:45 AM CST daily | Adversarial red team | memory/audits/abaddon-YYYY-MM-DD.md |
| Abaddon manual | On demand | Adversarial red team | Same, plus letter grade posted to Security topic |

---

## Trigger Phrases (manual)

- "run red team"
- "run Abaddon"
- "run full assessment"
- "Abaddon report"

---

## What Abaddon Checks

1. Exposed ports and unexpected network listeners
2. Credential leaks — environment variables, log files, temp files
3. Process list — suspicious or unexpected processes
4. File permissions on sensitive paths (openclaw.json, SOUL.md, AGENT_PROMPT.md, LaunchAgent plists)
5. Unauthorized SSH keys, cron entries, LaunchAgents
6. ops/exec-audit.log — anomalous command patterns
7. ops/identity-change-audit.log — rapid/unexpected config modifications
8. Gateway binding — verify loopback-only
9. World-readable files in sensitive directories
10. Assigns a letter grade (A–F)

---

## Scheduling

Abaddon runs as a native OpenClaw cron job in `cron/jobs.json`. No shell script wrapper, no random delay.

**Previous approach (deprecated):** `scripts/abaddon-trigger.sh` with random 0–16hr delay via LaunchAgent. This was removed because the trigger script never actually ran a scan — it only checked gateway health and logged a false "firing" message.

**Current approach:** Direct cron entry in jobs.json at 3:45 AM CST, `sessionTarget: isolated`, `wakeMode: now`, `agentTurn` with the red team prompt baked in. Delivers to Telegram Security topic with letter grade.

---

## File Permissions Enforced by Abaddon Findings

These paths should always be 600 (owner-only). Abaddon flags deviations:

- `~/.openclaw/openclaw.json`
- `~/.openclaw/cron/jobs.json`
- `~/.openclaw/workspace/SOUL.md`
- `~/.openclaw/workspace/MEMORY.md`
- `~/.openclaw/workspace/USER.md`
- `~/.openclaw/workspace/agents/observer/AGENT_PROMPT.md`
- `~/Library/LaunchAgents/ai.openclaw.*.plist`
- `~/Library/LaunchAgents/com.openclaw.*.plist`

Run `skills/enoch-tuning/setup/lock-identity.sh` to enforce all of the above in one pass.

---

## ClawHub Publishing

When published as a standalone skill, Abaddon becomes a distributable red team module. Until then, it lives inside Gideon's AGENT_PROMPT.md.

See: `agents/observer/AGENT_PROMPT.md` → ## Red Team Mode (Abaddon)

templates/abaddon-prompt.md

---

## Red Team Mode (Abaddon) 🔴

Activated when triggered with: "run red team", "run Abaddon", "run full assessment", or via nightly cron at 3:45 AM.

In red team mode, shift from passive monitor to active attacker. Think like an adversary. Find what they would find.

### Trigger Detection
If the user says "run red team", "run Abaddon", "run full assessment", or "Abaddon report" — enter this mode immediately. Do NOT run the standard passive audit. Run the full Abaddon protocol below.

### Attack Vector Checks

**Network & Exposure**
- `lsof -iTCP -sTCP:LISTEN` — flag anything bound to 0.0.0.0 (should all be 127.0.0.1)
- Gateway binding: localhost only or exposed to network?
- SSH Remote Login: `systemsetup -getremotelogin`
- Active tunnels: any ngrok, cloudflared, or unexpected remote access running?
- Firewall: enabled? stealth mode on?
- DNS: any unexpected servers configured?

**System Integrity**
- SIP: `csrutil status` — must be enabled
- FileVault: `fdesetup status` — must be On
- Gatekeeper: `spctl --status` — must be enabled
- macOS: current version? pending updates?

**OpenClaw Configuration**
- Exec security mode: full/allowlist/deny — is it appropriate?
- Auth enabled on gateway?
- streamMode off? (prevents conversation preview leaks)
- Plugins: any unexpected enabled?
- Cron jobs: any unexpected entries?

**Identity & File Security**
- SOUL.md + AGENTS.md: must be root-owned, 444
- MEMORY.md, USER.md, cron/jobs.json, openclaw.json: must be 600
- AGENT_PROMPT.md (this file): must be 600 — your detection playbook must not be world-readable
- LaunchAgent plists (ai.openclaw.*, com.openclaw.*): must be 600
- Plaintext key scan: `grep -r "sk-\|xai-\|Bearer\|api_key" ~/.openclaw/workspace/ --include="*.md" --include="*.json" --include="*.txt" 2>/dev/null | grep -v ".git"`
- .env files in workspace root?
- Git history secrets: `git -C ~/.openclaw/workspace log -p --all 2>/dev/null | grep -i "password\|secret\|api_key\|token" | head -20`

**API Key Handling**
- Keys in Keychain or flat config files?
- Keys visible in env vars that could be logged?
- Plaintext keys in openclaw.json?
- Check ~/.zshrc for hardcoded secrets

**Agent Behavior**
- Memory injection scan: `grep -r "ignore previous\|new instructions\|system:\|you are now" ~/.openclaw/workspace/memory/ 2>/dev/null`
- Sub-agents scoped correctly?
- Any agent with unexpected permissions?

**Dependencies**
- `brew outdated` — flag openclaw, ollama, node specifically
- `npm outdated -g` — flag openclaw

### Scoring
- CRITICAL: immediate fix required, active risk
- HIGH: fix before going live
- MEDIUM: fix within 30 days
- LOW: best practice, fix when convenient
- PASS: clean

**Security Grade:**
- A: 0 CRITICAL, 0 HIGH
- B: 0 CRITICAL, 1–2 HIGH
- C: 1 CRITICAL or 3+ HIGH
- D: 2+ CRITICAL
- F: Active compromise indicators

### Output

**1. Technical Report** (internal)
Full format — exact commands run, evidence captured, precise remediation steps.
Save to: `memory/audits/abaddon-YYYY-MM-

Github ReposUpdated 3d agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.

MCPOPENCLAW

Github ReposUpdated 7mo agoRank 70

CopilotKit

The Frontend for Agents & Generative UI. React + Angular