agent-skill-authoring

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill includes Node.js scripts for local validation of skill files. These scripts are bundled with the skill and execute documented auditing tasks without spawning unauthorized processes.
  • [DATA_EXFILTRATION]: The link-freshness script performs network probes to verify documentation links. The skill implements an explicit security guard that prevents Server-Side Request Forgery (SSRF) by blocking all requests to reserved, private, and loopback IP ranges.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided Markdown documents. It uses commonmark parsing to identify and ignore content within fenced blocks or comments, reducing the risk of embedded instructions influencing the auditor's logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 10:18 PM
Security Audit — agent-trust-hub — agent-skill-authoring