The Agent Skills Directory

[SAFE]: The skill is a specialized tool for formal methods maintenance (TLA+, Alloy, Z3, etc.). The instructions focus on mapping business rules to formal properties and implementation code.
[SAFE]: Evaluation files (evals/*.yaml) and fixtures (evals/fixtures/*) are part of a self-testing framework and do not contain executable malicious code. They use synthetic logs and mock models to verify the agent's reasoning capabilities.
[SAFE]: The README.md and SKILL.md documents promote best practices such as least-privilege reporting and domain-language translation of technical failures.
[SAFE]: All external references (such as agentskills.io and GitHub links to mizchi/skills) are consistent with the author's identity and provided context.
[SAFE]: No obfuscation, prompt injection, or persistence mechanisms were found across the 21 analyzed files.

formal-methods-drift-guard