The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The documentation provides instructions to install companion skills using npx skills add crissmoldovan/agent-skills. This involves downloading and executing code from the author's public package repository.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute various shell commands for repository inspection, including git grep, git branch, and git log. These commands are used to enumerate callers, identify work in flight, and verify repository state.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local repository (code, schema files, commit history), which constitutes an attack surface for indirect prompt injection.
Ingestion points: Reads file contents via the Read tool and repository metadata through git commands.
Boundary markers: The skill instructions emphasize restating the user brief as a change set and using 'control' searches to verify empty results, which serves as a logical check on ingested data.
Capability inventory: The skill uses Bash for command execution, Read for file access, and Write for generating 'run records' in the docs/ directory.
Sanitization: No explicit sanitization of codebase content is mentioned before it is interpolated into the final analysis output.

blast-area