agent-skill-offering

Pass

Audited by Gen Agent Trust Hub on Oct 27, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides best-practice guidance for versioning and distributing agent skills, emphasizing the 'lockstep rule' to ensure documentation matches the tool version.
  • [SAFE]: The included Python script scripts/check-skill-version.py uses only standard library modules (argparse, json, re, sys, pathlib) and performs local file operations on metadata files without performing any network access or downloading external code.
  • [SAFE]: The documentation in references/ci-release-gates.md correctly identifies security risks such as remote code execution and metadata poisoning, recommending the use of reputable security scanners like Snyk to mitigate these risks.
  • [SAFE]: Analysis of all 11 threat categories, including prompt injection, obfuscation, and data exfiltration, yielded no positive detections in the skill's instructions, scripts, or examples.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 27, 2026, 07:20 AM