release-installer

Pass

Audited by Gen Agent Trust Hub on Nov 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a custom Node.js script (release.mjs) to execute shell commands for repository operations, package management, and interactions with the GitHub API via the gh CLI tool.
  • [EXTERNAL_DOWNLOADS]: During the testing and verification phases, the skill downloads external resources including the semver package for version logic and the @earendil-works/pi-coding-agent package for integration testing via the npm registry.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as git commit messages and project metadata to determine the appropriate version bump for a release.
  • Ingestion points: Git log history, package.json, and mcp.pi.json (SKILL.md, release.mjs).
  • Boundary markers: Absent from the data parsing logic, but the skill explicitly instructs the agent to seek user confirmation before acting on suggestions.
  • Capability inventory: Includes creating git tags, pushing to origin, and creating GitHub Pull Requests.
  • Sanitization: Version strings are validated against a strict semantic versioning regular expression pattern.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 17, 2026, 04:04 PM