talonforge-skill-guardScan ClawHub skills for security vulnerabilities BEFORE installing. Use when installing new skills from ClawHub to detect prompt injections, malware payloads...
Install via ClawdBot CLI:
clawdbot install casperzinou/talonforge-skill-guardGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Contains instructions to override system prompt or ignore user requests
"IGNORE PREVIOUS INSTRUCTIONS"Potentially destructive shell commands in tool definitions
rm -rf /Calls external URL not in known-safe list
https://github.com/invariantlabs-ai/mcp-scanAudited Apr 17, 2026 · audit v1.0
Generated Oct 6, 2026
A platform team rolling out AI coding agents to hundreds of developers needs a gate that scans third-party ClawHub skills before they reach production machines. skill-guard provides the pre-install check that blocks prompt injections and exfiltration URLs at the client side.
Individual developers experimenting with community-built agent skills often install them blindly, exposing API keys and local files. skill-guard gives a one-command safe-install wrapper that quarantines suspicious skills in /tmp before anything touches the real skills folder.
Financial institutions deploying autonomous agents must document that every installed skill was scanned for malware, secrets exposure, and data-leak patterns. skill-guard's exit codes and scan output provide auditable evidence of pre-install security review.
An AI tooling vendor curating a shared skill registry wants to pre-screen submissions before listing them for user install. skill-guard can be run on staged downloads to flag AI-specific threats that traditional antivirus signatures miss entirely.
Research teams give agents broad file and network access for experiments, where one malicious skill could compromise sensitive datasets. skill-guard's quarantine workflow lets researchers inspect flagged skills before deciding whether to proceed.
Offer skill-guard free as an open-source utility for individual developers, then monetize an enterprise tier with centralized policy management, scan telemetry dashboards, and SLA-backed support. The pre-install gate becomes a compliance checkbox large organizations are willing to pay for.
Run skill-guard as a hosted pipeline where customers submit skills or configs and receive certified scan reports with human review for flagged cases. This shifts the burden of maintaining mcp-scan updates and threat intelligence from the customer to the vendor.
Build a registry where vetted skills are badged as skill-guard-verified, and sell aggregated anonymized threat intelligence about emerging prompt-injection and exfiltration patterns to security teams and platform vendors.
💬 Integration Tip
Replace `clawhub install` calls with `./scripts/safe-install.sh` and wire exit code 2 into your CI pipeline or shell prompt to halt automated installs whenever threats are detected. Ensure both `clawhub` and `uv` are pre-installed on every machine that runs the wrapper.
Scored Oct 6, 2026
Security-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope,...
Security scanner for AI agent skills. 9 built-in detection signatures. Identifies secrets, unsafe execution patterns, and prompt injection. Sub-50ms results.
Wallet anti-theft guard. One-click scan for high-risk wallet approvals to protect user assets. Use when a user asks for a wallet security check, wallet healt...
Comprehensive security audit for an agent's full skill stack. Chains scanner, differ, trust-verifier, and health-monitor into a single assessment with priori...
GEO Audit — AI Search Visibility Checker for ChatGPT, Perplexity, Claude & Gemini. 29-point GEO readiness checklist: robots.txt AI crawler access, Index...
Senior SecOps engineer skill for application security, vulnerability management, compliance verification, and secure development practices. Runs SAST/DAST sc...