Marketplace Skill Risk Scanner Pro v1.0

Version: 1.0.0

Released: 2026-05-04

Last security re-audit: 2026-05-07

Price: $4.99

Category: security

QA Status: Approved for release; security-cleared (re-confirmed 2026-05-07)

Overview

Marketplace Skill Risk Scanner Pro is the pre-install vetting tool for serious OpenClaw operators. Deep static analysis flags remote script execution, destructive commands, embedded secrets, dynamic code execution, privilege escalation, and network exfiltration — before untrusted code ever touches your stack. Outputs a ranked severity report with evidence snippets (secrets redacted), risk rating (LOW/MEDIUM/HIGH/CRITICAL), and an install recommendation. Never executes inspected code. Essential for anyone installing third-party or marketplace skills in a production environment.

What you get

OpenClaw skill package
SKILL.md usage guide
scripts/scan_skill_risk.py — stdlib-only static analysis engine
Markdown and JSON output modes

Usage


# Markdown report (default)
python3 scripts/scan_skill_risk.py /path/to/skill-dir

# JSON output for pipelines
python3 scripts/scan_skill_risk.py /path/to/skill-dir --format json

# Limit scan to files under 256 KB
python3 scripts/scan_skill_risk.py /path/to/skill-dir --max-file-kb 256

Output

Risk rating: LOW / MEDIUM / HIGH / CRITICAL
Findings: suspicious patterns by file and line with evidence snippets (secrets redacted)
Metadata gaps: missing ClawHub metadata, integration-fit issues
Install recommendation: ALLOW / REVIEW / REJECT

Security Boundary

No execution of inspected code
SENSITIVE_NAMES exclusion guard prevents scanning MEMORY.md, USER.md, SOUL.md, .env, openclaw.json, and credential files
--allow-sensitive-paths opt-in required for any sensitive path override
All RISK_PATTERNS regex lines tagged # nosec: pattern-definition — not shell strings
No subprocess, no network calls, no eval/exec
stdlib-only (re, pathlib, argparse, json — no pip installs)

Security & QA

Viper: APPROVED_FOR_RELEASE (5/5 functional tests passed — 2026-05-04)
Warlock: SECURITY_CLEARED (2026-05-05) — sensitive file exclusion + allow-sensitive-paths opt-in
Warlock re-audit: SECURITY_CLEARED (2026-05-07) — all 14 audit findings confirmed false positives: SENSITIVE_NAMES is exclusion guard; re.compile() calls are pattern definitions (nosec respected); $HOME in regex literal not shell variable; skill-security-audit CLEAN; py_compile PASS; functional test PASS (risk_score=3/allow on clean target)

Requirements

Python 3.8+ (stdlib only — no pip installs)
OpenClaw v2026.3.23 or later
No API keys required

Marketplace Skill Risk Scanner Pro