Security Audit Tianjin

Comprehensive security auditing for Clawdbot deployments. Scans for exposed credentials, open ports, weak configs, and vulnerabilities. Auto-fix mode included.

Security Scan

The skill mostly matches a Clawdbot security-audit purpose, but there are inconsistencies (missing declared runtime requirement for node, hard-coded root paths, and unverified auto-fix behavior) that warrant caution before running it.

Purpose & Capability

The name/description (Clawdbot security audit) aligns with the script's behaviors: recursive file scans, port checks, Docker/Git checks and credential pattern matching under a Clawdbot directory. However, the script hard-codes paths under /root/clawd and a CONFIG_DIR '/root/clawd/skills/.env' which is very specific and could be incorrect or overly broad for other installs. Also _meta.json ownerId differs from registry ownerId (metadata mismatch).

Instruction Scope

SKILL.md instructs to run node skills/security-audit/scripts/audit.cjs with flags and promises scanning of 'tokens in command history' and an 'auto-fix' mode. The included script shows many read/scan operations and execSync usage (ss/netstat, git), but the provided file is truncated so the full auto-fix implementation and any reads of shell history are not visible. The instructions give the agent permission to run code that reads many files and may modify filesystem state; the exact modification behavior (what --fix changes) is not fully verifiable from the truncated code.

Install Mechanism

There is no install spec (instruction-only style), but SKILL.md commands invoke node. The registry metadata lists no required binaries. That is an inconsistency: the script requires Node.js to run but the skill does not declare Node as a required binary or runtime. Running without Node will fail; running with Node will execute the bundled code. Lack of declared runtime is a mismatch the user should note.

Credentials

The skill declares no environment variables or external credentials (consistent). The script reads many files under /root/clawd (configs, .env, Dockerfile, .git, code files) which is proportionate to auditing Clawdbot, but because the path is hard-coded to /root it will access root-owned files if run as a privileged user. The audit's read-only checks are expected, but the auto-fix option implies write operations (e.g., setting permissions, creating .gitignore) which require filesystem write privileges and increase risk if run on a system with unrelated sensitive files.

Persistence & Privilege

The skill is not always-enabled and the default autonomous invocation is allowed (platform default). There are no indications it attempts to persist itself or modify other skills. However, the auto-fix behavior can change file permissions and create files in the audited tree — this is expected for an auto-fix feature but is higher privilege than a pure scanner and should be used cautiously.

What to consider before installing

This skill appears to implement a Clawdbot security audit, but take these precautions before installing/running it: - Expect to run the script with Node.js; the skill does not declare Node as a required binary — ensure Node is present or the run will fail. - Review the full scripts/audit.cjs source (the provided file was truncated) to confirm exactly what --fix does (which files are modified, exact permission changes, any writes). Back up the target directories first. - The script uses hard-coded paths under /root/clawd and '/root/clawd/skills/.env'. Only run it on a system where those paths are the intended target; otherwise edit the script or run it in a controlled environment. - Run the audit in read-only mode (no --fix) first to inspect findings. If you consider using --fix, run it in an isolated test environment or container and review changes afterwards. - The code uses child_process.execSync (ss/netstat, git). That is expected for port/git checks but increases the blast radius — avoid running as root unless necessary. - Confirm the skill source/owner is trusted (metadata mismatch observed). If provenance is uncertain, prefer manual review or use in an isolated sandbox.

Security Audit Skill

When to use

Run a security audit to identify vulnerabilities in your Clawdbot setup before deployment or on a schedule. Use auto-fix to remediate common issues automatically.

Setup

No external dependencies required. Uses native system tools where available.

How to

Quick audit (common issues)

node skills/security-audit/scripts/audit.cjs

Full audit (comprehensive scan)

node skills/security-audit/scripts/audit.cjs --full

Auto-fix common issues

node skills/security-audit/scripts/audit.cjs --fix

Audit specific areas

node skills/security-audit/scripts/audit.cjs --credentials      # Check for exposed API keys
node skills/security-audit/scripts/audit.cjs --ports            # Scan for open ports
node skills/security-audit/scripts/audit.cjs --configs          # Validate configuration
node skills/security-audit/scripts/audit.cjs --permissions      # Check file permissions
node skills/security-audit/scripts/audit.cjs --docker           # Docker security checks

Generate report

node skills/security-audit/scripts/audit.cjs --full --json > audit-report.json

Output

The audit produces a report with:

Level	Description
🔴 CRITICAL	Immediate action required (exposed credentials)
🟠 HIGH	Significant risk, fix soon
🟡 MEDIUM	Moderate concern
🟢 INFO	FYI, no action needed

Checks Performed

Credentials

• API keys in environment files
• Tokens in command history
• Hardcoded secrets in code
• Weak password patterns

Ports

• Unexpected open ports
• Services exposed to internet
• Missing firewall rules

Configs

• Missing rate limiting
• Disabled authentication
• Default credentials
• Open CORS policies

Files

• World-readable files
• Executable by anyone
• Sensitive files in public dirs

Docker

• Privileged containers
• Missing resource limits
• Root user in container

Auto-Fix

The --fix option automatically:

• Sets restrictive file permissions (600 on .env)
• Secures sensitive configuration files
• Creates .gitignore if missing
• Enables basic security headers

Related skills

• security-monitor - Real-time monitoring (available separately)