share-a-library

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes git, gh, and npx shell commands to initialize a repository, build documentation, and publish the workspace to GitHub.
  • [DATA_EXFILTRATION]: The workflow performs git add . followed by gh repo create --public --push, which publishes the entire content of the current directory to a public GitHub repository. If the workspace contains sensitive files such as .env files, API keys, or private SSH keys that are not explicitly ignored, they will be exfiltrated and publicly exposed.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill ingests all files within the local workspace directory (SKILL.md).
  • Boundary markers: None. The skill does not provide instructions to the agent to filter or verify files before publishing.
  • Capability inventory: Command execution (git, gh, npx) and public network exfiltration (GitHub push).
  • Sanitization: None. The tool lacks validation steps to ensure sensitive data is not being published.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 03:32 AM
Security Audit — agent-trust-hub — share-a-library