security-audit
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary purpose is to audit untrusted code and configurations, which involves ingesting potentially adversarial data.
- Ingestion points: Project source files (scanned via
scripts/scanners/*.sh) and repository metadata (fetched viascripts/github-security-audit.sh). - Boundary markers: Explicit boundary instructions are absent from the
SKILL.mdprompt, although they are recommended in the reference documentation (references/llm-security.md). - Capability inventory: The skill's tool access is restricted to read-only operations and diagnostic utilities (
grep,jq,gh). It lacks capabilities for writing files or executing arbitrary shell commands. - Sanitization: None; the tool reports raw matches from the scanned data.
- [SAFE]: The skill performs its intended function using standard command-line tools. All identified patterns from static analysis hints (e.g., destructive commands or instruction overrides) are located within educational reference material or provided as secure examples and do not constitute a threat.
Audit Metadata