The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub issues, PR comments, and commit messages during the research phase. Malicious instructions embedded in these external sources could potentially influence the agent's behavior during the synthesis of the final document.
Ingestion points: git clone, gh pr view, and gh issue list operations fetch content from the public React repository and GitHub discussions.
Boundary markers: The skill does not explicitly define delimiters or 'ignore' warnings for the data being synthesized into the final report.
Capability inventory: The skill has file-write capabilities to the .claude/research/ directory and utilizes subprocess calls via the gh and git CLIs.
Sanitization: No explicit sanitization or filtering of the fetched external content is described before interpolation into the final document.
[EXTERNAL_DOWNLOADS]: The skill clones the React repository from GitHub (https://github.com/react/react.git) and uses the GitHub CLI to fetch PR and issue data. These operations target well-known and authoritative services.
[COMMAND_EXECUTION]: The skill executes various CLI commands including git clone, git pull, git log, gh pr list, and gh issue list. These commands are standard for the skill's intended purpose of technical research.

react-expert