The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes system tools such as gh, npm, grep, and python to manage its issue registry. It implements high-quality security hygiene by using single-quoted heredocs (e.g., cat <<'ISSUE_BODY') in context/action-create.md to ensure that user-provided content in issue bodies cannot be executed as shell commands. Additionally, the registry_dir configuration includes strict validation against path traversal, rejecting absolute paths, parent directory segments (..), and symlinks that resolve outside the project directory.
[EXTERNAL_DOWNLOADS]: The skill fetches data from trusted and well-known services, specifically GitHub (github.com) and Anthropic's official status page (status.claude.com), as well as Marginlab (marginlab.ai). These operations are essential for tracking product bugs and service health and align with the skill's primary purpose.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because its core function involves reading and displaying untrusted data from external sources (GitHub issue titles, descriptions, and community-reported status updates).
Ingestion points: Data enters the agent's context through gh issue view, gh search, and web requests to external status pages.
Boundary markers: Output templates in context/output-templates.md lack explicit boundary markers or instructions for the agent to ignore embedded commands in the external data.
Capability inventory: The skill has the capability to execute shell commands, write to the local file system (registry management), and create new GitHub issues.
Sanitization: Mitigation is present in the form of schema validation in scripts/registry_manager.py and safe shell quoting in the provided bash scripts.

known-issues