release-skills

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple local shell commands for version control and repository management.
  • Uses git for commit analysis, tagging, and pushing changes.
  • Uses the GitHub CLI (gh) to view repository/PR information and create/edit releases.
  • Implements safety guardrails by using --notes-file for release notes, which prevents potential command injection that could occur if multiline changelog content were passed directly as string arguments.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources that could contain untrusted input.
  • Ingests commit messages from Git history and PR author information from the GitHub API.
  • While this content is used to populate changelogs and release notes, the skill incorporates boundary markers by instructing the agent to write these notes to temporary files rather than processing them directly in shell execution strings.
  • [DYNAMIC_EXECUTION]: Supports project-specific execution of shell hooks defined in the project's repository.
  • The skill can trigger prepare_artifact and publish_artifact hooks defined in a local .releaserc.yml file.
  • These hooks are intended for repository-specific build and deployment logic, representing a standard extensibility pattern for development tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 07:47 AM
Security Audit — agent-trust-hub — release-skills