release-notes-narrative

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands including git, gh, jq, and mkdir to manage local context and interact with the GitHub API for release management.
  • [EXTERNAL_DOWNLOADS]: Fetches external data from GitHub using the gh pr view command to incorporate pull request descriptions into release notes.
  • [PROMPT_INJECTION]: The skill is exposed to indirect prompt injection as it processes text from commit messages and pull request bodies which are controlled by potentially multiple contributors.
  • Ingestion points: Ingests data through git log (Step 4) and gh pr view (Step 6) in SKILL.md.
  • Boundary markers: There are no explicit instructions to the AI to treat the ingested commit or PR text as untrusted data or to use specific delimiters during the synthesis phase.
  • Capability inventory: The skill possesses the capability to modify repository releases using gh release edit (Step 12) and manipulate local files via shell redirection.
  • Sanitization: Employs a markdown AST-based sanitization step to remove problematic autolinks (Step 10) and a preflight validation check (Step 11) before applying changes.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 07:58 PM
Security Audit — agent-trust-hub — release-notes-narrative