The Agent Skills Directory

[SAFE]: The skill is configured with a restricted set of tools (Read, Grep, Glob, Bash(ls:*), Bash(find:*), Bash(wc:*), Agent) which are appropriate for its stated purpose of analyzing local files and source code for test coverage.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests untrusted data from documentation and source code files provided via user arguments.
Ingestion points: Feature documentation and source code paths provided through $ARGUMENTS in SKILL.md.
Boundary markers: Not present.
Capability inventory: Read, Grep, Glob, Bash (ls, find, wc), and Agent tool for subagent dispatch.
Sanitization: No sanitization is applied to the content of the read files.
Note: The risk is assessed as safe due to the highly restricted nature of the available shell commands and the absence of network or file-write capabilities.

check-coverage