The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/agent.py uses subprocess.run to invoke the gh (GitHub) CLI for various API tasks. This is implemented using a list-based argument structure, which is a standard secure practice for CLI interactions that prevents shell injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the GitHub API when retrieving security alerts and repository information in scripts/agent.py and scripts/process.py.\n
Ingestion points: Data is fetched from GitHub's Code Scanning and Secret Scanning API endpoints.\n
Boundary markers: The processing logic does not employ specific delimiters or boundary instructions to isolate external data from the execution environment.\n
Capability inventory: The skill has permissions to write to the local file system (generating JSON reports) and conduct network operations via the GitHub CLI and urllib.\n
Sanitization: Although alert descriptions are truncated for display in summaries, no formal sanitization is performed. However, because this data is used for reporting rather than further prompt interpolation or instruction following, the risk is minimal.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to the official GitHub API (api.github.com) to fetch security metadata and alert details. As GitHub is a well-known and trusted service within the development ecosystem, these operations are documented neutrally and considered safe for the skill's intended administrative purpose.

implementing-github-advanced-security-for-code-scanning