Clawdbot-Security-Check - GitLite

KO | EN

gitlite — search

#c #postgresql #json #mysql #gemini #generative-ai #animation #claude #svg #editor

Clawdbot-Security-Check

(★ 63)

This skill teaches Clawdbot to audit its own security posture using first-principles reasoning. Not a hard-coded script—it's a knowledge framework that Clawdbot applies dynamically to detect vulnerabilities, understand their impact, and recommend specific remediations.

파일 탐색기

ZIP 다운로드

.gitignore
README.md
skill.json
SKILL.md

# CDN으로 사용하기

jsDelivr

jsDelivr는 공개 GitHub 리포지토리를 별도 설정 없이 CDN으로 즉시 서빙합니다. 버전과 파일을 고르면 웹페이지에 바로 붙일 수 있는 링크와 예시 코드가 만들어집니다.

버전

파일

링크

예시

명령어 용어집

이 문서에서 사용된 명령어를 모아봤습니다. 낯선 명령어가 있다면 펼쳐서 확인해보세요.

🔍

`git clone`

설명 보기 ▼

Clone an existing repository.

git clone {{remote_repository_location}} {{path/to/directory}}

Clone an existing repository into a new directory (the default directory is the repository name):

git clone --recursive {{remote_repository_location}}

Clone an existing repository and its submodules:

git clone {{[-n|--no-checkout]}} {{remote_repository_location}}

Clone only the `.git` directory of an existing repository:

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

🔒 **Self-security audit framework for Clawdbot**

Inspired by the security hardening framework from [ᴅᴀɴɪᴇʟ ᴍɪᴇssʟᴇʀ](https://x.com/DanielMiessler) and integrated with [official ClawdBot security documentation](https://docs.clawd.bot/gateway/security).

This skill teaches Clawdbot to audit its own security posture using first-principles reasoning. Not a hard-coded script—it's a **knowledge framework** that Clawdbot applies dynamically to detect vulnerabilities, understand their impact, and recommend specific remediations.

## What This Is

- 🧠 **Knowledge-based** - Embeds the security framework directly in Clawdbot
- 🔍 **Dynamic detection** - Clawdbot learns to find issues, not just run a script
- 📚 **Extensible** - Add new checks by updating the skill
- 🔒 **100% Read-only** - Only audits; never modifies configuration
- 📖 **Report-focused** - Generates detailed reports with remediation guidance

## The 13 Security Domains

| # | Domain | Severity | Key Question |
|---|--------|----------|--------------|
| 1 | Gateway Exposure | 🔴 Critical | Is the gateway bound to 0.0.0.0 without auth? |
| 2 | DM Policy | 🟠 High | Are DMs restricted to an allowlist? |
| 3 | Group Access Control | 🟠 High | Are group policies set to allowlist? |
| 4 | Credentials Security | 🔴 Critical | Are secrets in plaintext with loose permissions? |
| 5 | Browser Control Exposure | 🟠 High | Is remote browser control secured? |
| 6 | Gateway Bind & Network | 🟠 High | Is network exposure intentional and controlled? |
| 7 | Tool Access & Sandboxing | 🟡 Medium | Are tools sandboxed with minimal access? |
| 8 | File Permissions & Disk | 🟡 Medium | Are file permissions properly set? |
| 9 | Plugin Trust & Model | 🟡 Medium | Are plugins allowlisted and models current? |
| 10| Logging & Redaction | 🟡 Medium | Is sensitive data redacted in logs? |
| 11| Prompt Injection | 🟡 Medium | Is untrusted content wrapped? |
| 12| Dangerous Commands | 🟡 Medium | Are destructive commands blocked? |
| 13| Secret Scanning | 🟡 Medium | Is detect-secrets baseline maintained? |

## Installation

```bash
# Via ClawdHub
clawdhub install clawdbot-security-check

# Or clone for manual installation
git clone https://github.com/TheSethRose/Clawdbot-Security-Check.git
cp -r Clawdbot-Security-Check ~/.clawdbot/skills/
```

## Usage

### Via Clawdbot
```
@clawdbot audit my security
@clawdbot run security check
@clawdbot what vulnerabilities do I have?
@clawdbot security audit --deep
```

**Note:** The `--fix` flag is documented for reference but the skill itself only generates reports—it never automatically modifies configuration.

## Security Principles

Running an AI agent with shell access requires caution. Focus on three areas:

1. **Who can talk to the bot** — DM policies, group allowlists, channel restrictions
2. **Where the bot is allowed to act** — Network exposure, gateway binding, proxy configs
3. **What the bot can touch** — Tool access, file permissions, credential storage

### Trust Hierarchy

Apply appropriate trust levels based on role:

| Level | Entity | Trust Model |
|-------|--------|-------------|
| 1 | **Owner** | Full trust — has all access |
| 2 | **AI** | Trust but verify — sandboxed, logged |
| 3 | **Allowlists** | Limited trust — only specified users |
| 4 | **Strangers** | No trust — blocked by default |

## High-Level Audit Checklist

Treat findings in this priority order:

1. 🔴 Lock down DMs and groups if tools are enabled on open settings
2. 🔴 Fix public network exposure immediately
3. 🟠 Secure browser control with tokens and HTTPS
4. 🟠 Correct file permissions for credentials and config
5. 🟡 Only load trusted plugins
6. 🟡 Use modern models for bots with tool access

## Key Features

### Prompt Injection Mitigation
- Keep DMs locked to `pairing` or `allowlists`
- Use mention gating in groups
- Treat all links and attachments as hostile
- Run sensitive tools in a sandbox
- Use instruction-hardened models like Anthropic Opus 4.5

### Sandboxing Levels
| Mode | Description |
|------|-------------|
| `none` | Workspace is off limits |
| `ro` | Workspace mounted read-only |
| `rw` | Workspace mounted read-write |

### Incident Response (if compromised)
1. **Containment** — Stop gateway, set bind to loopback, disable risky DMs/groups
2. **Rotation** — Change gateway token, rotate browser/hook tokens, revoke API keys
3. **Review** — Check logs, review config changes, re-run audit with `--deep`

### Secret Scanning (detect-secrets)
```bash
# Find candidates
detect-secrets scan --baseline .secrets.baseline

# Review findings
detect-secrets audit

# Update baseline after rotating secrets
detect-secrets scan --baseline .secrets.baseline --update
```

## Extending the Framework

Add new checks by contributing to SKILL.md:

```markdown
## 14. New Vulnerability 🟡 Medium

**What to check:** What config reveals this?

**Detection:**
```bash
command-to-check-config
```

**Vulnerability:** What can go wrong?

**Remediation:**
```json
{ "fix": "here" }
```
```

## Architecture

```
Clawdbot-Security-Check/
├── SKILL.md      # Knowledge framework (the skill - source of truth)
├── skill.json    # Clawdbot metadata
├── README.md     # This file
└── .gitignore
```

**SKILL.md** is the source of truth—it teaches Clawdbot everything it needs to know.

## Why This Approach?

Hard-coded scripts get stale. A knowledge framework evolves:

- ✅ Add new vulnerabilities without code changes
- ✅ Customize checks for your environment
- ✅ Clawdbot understands the "why" behind each check
- ✅ Enables intelligent follow-up questions

> "The goal isn't to find vulnerabilities—it's to understand security deeply enough that vulnerabilities can't hide." — Daniel Miessler

## Output Example

```
═══════════════════════════════════════════════════════════════
🔒 CLAWDBOT SECURITY AUDIT
═══════════════════════════════════════════════════════════════
Timestamp: 2026-01-26T15:30:00.000Z

┌─ SUMMARY ───────────────────────────────────────────────
│ 🔴 Critical:  1
│ 🟠 High:      2
│ 🟡 Medium:    1
│ ✅ Passed:    9
└────────────────────────────────────────────────────────

┌─ FINDINGS ──────────────────────────────────────────────
│ 🔴 [CRITICAL] Gateway Exposure
│    Finding: Gateway bound to 0.0.0.0:18789 without auth
│    → Fix: Set gateway.auth_token environment variable
│
│ 🟠 [HIGH] DM Policy
│    Finding: dm_policy is "allow" (all users)
│    → Fix: Set dm_policy to "allowlist" with trusted users
└────────────────────────────────────────────────────────

This audit was performed by Clawdbot's self-security framework.
No changes were made to your configuration.
```

## Reporting Vulnerabilities

Report security issues to: **security@clawd.bot**

**Do not post vulnerabilities publicly** until they have been fixed.

## Contributing

1. Fork the repo
2. Add new security knowledge to SKILL.md
3. Submit PR

## License

MIT - Security-first, open source forever.

---

**Clawdbot knows its attack surface. Do you?**