The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill is designed as a metadata and instruction linter for development workflows.
[PROMPT_INJECTION]: The static analysis tool flagged phrases such as "ignore previous instructions" in the marketplace-checks.md file. Detailed review confirms these are part of a checklist used by the linter to help developers detect and prevent prompt injection in their own skills; they are not instructions for the agent to follow.
[REMOTE_CODE_EXECUTION]: The documentation files (.beads/README.md and cli.md) provide instructions for users to manually install external tools like the "Beads" issue tracker and "skillcheck-cli" using curl | bash or go install. These are standard installation procedures for developers and are not executed by the agent skill itself.

skill-check