The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a local utility script, find-polluter.sh, which automates running tests to detect environment pollution. It uses standard commands like find and npm test locally to assist developers during debugging.\n- [INDIRECT_PROMPT_INJECTION]: The methodology requires the agent to ingest external technical data such as error messages and stack traces. While this is an ingestion surface, the skill's specific focus on multi-phase validation and root-cause evidence serves as a procedural safeguard against being misled by single data points.\n- [SAFE]: The skill promotes secure engineering practices, including defense-in-depth and environment guards to prevent accidental side effects during testing. No obfuscation, data exfiltration, or persistence mechanisms were detected.

systematic-debugging