The Agent Skills Directory

[SAFE]: The skill implements a robust, evidence-based auditing protocol that promotes transparency and rigorous verification of software development tasks. It uses standard markdown artifacts (audit tables and completion reports) to track progress.
[COMMAND_EXECUTION]: The skill instructs the agent to use standard development tools (e.g., git, pytest, npm test, gh CLI) to verify the state of the project. These commands are executed locally for the purpose of auditing and fixing code, which is the skill's primary function.
[EXTERNAL_DOWNLOADS]: The skill references standard package managers (e.g., npm, cargo, pip) and official tool registries for legitimate development tasks. These resources originate from well-known services and organizations.
[PROMPT_INJECTION]: The skill includes a detailed 'Rationalizations' section designed to prevent the AI from bypassing auditing rules or defaulting to unreliable 'Assumed Complete' statuses. This serves as a safety mechanism to ensure the agent adheres to its instructions rather than an attempt to bypass system guards.

check-completion