The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads code from GitHub repositories to install new agent skills. It supports both a curated list from a trusted organization (OpenAI) and arbitrary user-provided repositories. The download logic in scripts/install-skill-from-github.py includes a _safe_extract_zip function to prevent path traversal (Zip-Slip) attacks by verifying that all extracted files remain within the intended destination directory.
[COMMAND_EXECUTION]: The scripts/install-skill-from-github.py script executes git commands via subprocess.run to perform sparse checkouts. The command arguments are passed as a list rather than a single string, which effectively prevents shell injection vulnerabilities.
[CREDENTIALS_UNSAFE]: The skill utilizes GITHUB_TOKEN or GH_TOKEN environment variables to authenticate GitHub API requests. The scripts/github_utils.py module includes logic to include this token in the Authorization header. Security risk is mitigated by domain validation in the installation script, ensuring credentials are only sent to GitHub infrastructure.
[PROMPT_INJECTION]: The skill processes and displays names of directories from external GitHub repositories during the listing process. While this represents an ingestion point for untrusted data, the risk is inherent to the listing functionality and is mitigated by the agent's internal safety filters.

skill-installer