The Agent Skills Directory

[SAFE]: The SKILL.md file defines a structured interview process designed to resolve technical ambiguity. It includes guardrails to prevent implementation before key terms or ownership boundaries are clear, which promotes safe engineering practices.
[SAFE]: The provided evaluation scripts (eval/evaluate.py and eval/run-eval.sh) are used to verify the skill's adherence to its defined rules. The script performs validation checks against a set of scenarios and mutation tests to ensure the AI's logic remains consistent. These scripts operate on local files and do not perform network requests or access sensitive data.
[SAFE]: The scenarios defined in eval/fixtures/scenarios.json provide benign examples of engineering clarification turns and do not contain any malicious patterns or injection attempts.
[SAFE]: No obfuscation, data exfiltration, or unauthorized command execution patterns were identified across any of the analyzed files.

clarify-work