skills/mx-space/core/release-core/Gen Agent Trust Hub

release-core

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The prepare-server.sh script invokes npx to download and execute the conventional-changelog-cli package from the NPM registry. This is a standard development practice for generating changelogs from commit history using a well-known utility.\n- [INDIRECT_PROMPT_INJECTION]: The skill generates release notes by processing commit messages, which represents an ingestion point for potentially untrusted data. \n
  • Ingestion points: Git commit history retrieved via git log in preflight.sh.\n
  • Boundary markers: None explicitly defined in the data ingestion scripts, though the SKILL.md provides selection logic for the agent.\n
  • Capability inventory: File system writes (apps/core/RELEASE_NOTES.md), git operations (git commit, git tag, git push), and remote publishing (pnpm publish).\n
  • Sanitization: None; the agent is instructed to rewrite commit messages into user-facing language.\n- [COMMAND_EXECUTION]: The provided scripts automate various shell commands including git, npm, pnpm, and gh. These commands are used as intended for managing project releases, checking CI status, and publishing artifacts.\n- [CREDENTIALS_UNSAFE]: The release scripts perform a routine check for active NPM sessions using npm whoami. This is a best-practice verification step to ensure the environment is correctly configured for publishing and does not involve the exfiltration of credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 03:24 PM