The Agent Skills Directory

[SAFE]: The skill implements an 'inverted burden of proof' philosophy for code maintenance, focusing on identifying unreachable or unnecessary code symbols.
[COMMAND_EXECUTION]: The skill utilizes standard development tools for its operations, including git diff for scope identification and language-specific test runners (e.g., npm test, pytest, cargo test) for verification after cleanup. These are legitimate uses within a developer's environment.
[PROMPT_INJECTION]: No evidence of prompt injection, role-play instructions, or safety bypasses was detected. The instructions are purely functional and oriented toward code auditing.
[DATA_EXFILTRATION]: There are no network operations, external URL fetches, or attempts to access sensitive credentials. The scope is limited to the local codebase and version control system.
[OBFUSCATION]: The skill uses clear, human-readable markdown and text-based workflows with no encoded strings or hidden characters.
[PERSISTENCE_MECHANISMS]: No attempts to modify shell profiles, cron jobs, or registry keys were found. The skill operates strictly on the source code level within a defined session.

prune