The Agent Skills Directory

[SAFE]: No security issues were identified. The skill's operations are consistent with its purpose of assisting in software design and planning.
[COMMAND_EXECUTION]: Uses the bash tool specifically to run the GitHub CLI (gh issue view) for fetching project context. This is a restricted and legitimate use of shell capabilities to understand task requirements.
[EXTERNAL_DOWNLOADS]: Employs WebSearch and WebFetch to retrieve technical documentation and best practices from the internet. These are standard operations for research.
[PROMPT_INJECTION]: The skill processes potentially untrusted data from external sources (Indirect Prompt Injection surface). Ingestion points: GitHub issue bodies and external web content fetched during the research phase. Boundary markers: None explicitly defined for interpolated data. Capability inventory: Includes Bash (for gh), WebFetch, Read, and WebSearch. Sanitization: None. The risk is safe because the skill is strictly constrained to planning only and prohibits the agent from writing or executing code based on its research, maintaining a human-in-the-loop requirement.

design-plan