Agent Security Regression Evidence Pack 20260820 v1.0

What it does

Agent Security Regression Evidence Pack helps teams verify that agent security regression evidence is complete before release or handoff. It checks baseline and candidate versions, threat model, authentication, authorization, input validation, session isolation, data boundary, dependency audit, restore point, rollback trigger, and approval owner evidence.

The skill gives release teams a focused security-evidence readiness report without testing live systems or changing files. It is advisory, deterministic, local-only, and read-only.

Use cases

• Security reviewers checking agent regression evidence before approval.
• Release teams preparing rollback and approval records for production changes.
• Operators who need a repeatable local gate for redacted security notes.

Requirements

• Python 3 with standard-library modules only.
• Redacted inline text or one explicit local notes file.
• No API keys required.
• Compatible with the standard OpenClaw skill layout and local script execution.

Example usage

python3 scripts/agent_security_regression_evidence_pack.py --text "baseline version documented; candidate version documented; threat model documented; authentication test documented; authorization test documented; input validation test documented; session isolation test documented; data boundary test documented; dependency audit documented; restore point documented; rollback trigger documented; approval owner documented" --json

Expected output: JSON with result: "PASS", regression evidence coverage, and mutated_files: []. Incomplete evidence returns REVIEW; high-risk evidence or very low coverage returns BLOCKED.

Price

$4.99