The Agent Skills Directory

[SAFE]: No indicators of prompt injection, data exfiltration, or malicious intent were found in the provided files.
[SAFE]: The skill is purely instructional and contains no executable code, external dependencies, or network-enabled operations.
[SAFE]: The test scenarios (test-authority.md, test-convenience.md, test-false-complexity.md, and test-thrashing.md) provide structured reasoning exercises for evaluating model performance without introducing security risks.
[SAFE]: Mentions of security-sensitive paths like auth/token_validator.py and .env.example are used strictly as illustrative examples within hypothetical scenarios and do not involve actual file system access.

escalation-governance