skills/sasjs/skills/sasjs-release/Gen Agent Trust Hub

sasjs-release

Pass

Audited by Gen Agent Trust Hub on Nov 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes conventional git commit messages to generate release notes and determine version bumps, which introduces a risk of indirect prompt injection from repository contributors. Any user with commit access could potentially influence the agent's behavior during the release process through instructions hidden in commit metadata.
  • Ingestion points: Git commit messages are consumed and parsed from the repository history to automate release documentation.
  • Boundary markers: No specific delimiters or safety instructions are defined in the skill to separate untrusted commit data from the agent's operational logic.
  • Capability inventory: The skill performs file system operations (reading config files), git operations (creating and pushing tags), and network operations via API calls to GitHub, GitLab, and Gitea for release creation and asset attachment.
  • Sanitization: There is no mention of sanitizing or escaping commit message content before it is interpolated into host release notes or processed by the agent.
  • [COMMAND_EXECUTION]: The instructions involve executing CLI tools including npx sasjs-release and npm publish to perform versioning, documentation, and artifact distribution.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes npx to download and execute the @sasjs/release tool at runtime. This package is owned by the skill author (sasjs), representing a standard vendor dependency for the intended release workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 17, 2026, 08:31 PM