The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/gh_issue.py executes the GitHub CLI (gh) using subprocess.run to perform operations like fetching tokens, editing issues, and executing GraphQL queries. These calls are executed using argument lists rather than shell strings, which prevents shell injection attacks.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from GitHub (issue titles and bodies) and either displays it to the user or saves it to local files. This content is untrusted and could contain instructions designed to manipulate the AI agent.
Ingestion points: Data is fetched via GhApi and the gh CLI in scripts/gh_issue.py (e.g., list_issues, view, dump_tree).
Boundary markers: The skill does not implement explicit boundary markers or safety instructions when rendering or saving issue content.
Capability inventory: The skill can execute shell commands (gh CLI), write to the local filesystem, and make network requests to the GitHub API.
Sanitization: Issue titles are sanitized using regular expressions before being used for local file and directory names in the dump_tree command.
[EXTERNAL_DOWNLOADS]: The Python script declares dependencies (ghapi, typer, rich, pyyaml) to be managed by uv run. These are well-known packages from the standard Python Package Index (PyPI).

github-issue