| name | security-hardening |
| description | Use this skill when improving safety, reducing hidden side effects, tightening execution boundaries, or auditing trust boundaries during the Python port. |
Security hardening must be behavior-aware.
Tasks:
- identify trust boundaries
- identify unsafe implicit state
- identify fail-open risks
- identify side-effect-heavy startup assumptions
- propose hardening changes that preserve user-visible behavior
Rules:
- do not use security as an excuse for semantic drift
- document each hardening change with:
- prior risk
- preserved invariant
- verification method