The Agent Skills Directory

[SAFE]: The skill operates as a verification critic for requirements against source code. Its behavior is restricted to reading local artifacts, searching through source files, and writing verification reports to the project's filesystem.
[COMMAND_EXECUTION]: While the skill lists the Bash tool in its allowed tools, its usage is constrained by instructions to only perform read-only and verification tasks such as Grep or basic file manipulation for report generation. No arbitrary or high-risk command execution patterns were found.
[DATA_EXFILTRATION]: No network-capable tools or patterns (like curl, wget, or http requests) are used. Data access is limited to the local project structure (docs/reverse/ and source code).

re-verify-requirements