SKILLEMALL.ai
GitHub ↗ RU Sign in

BC red-team-verifier-patrick-munro

Adversarial verification of AI-generated legal content with systematic fact-checking, source validation, and quality control. Use when a user asks to verify, fact-check, red-team, validate sources, or quality-control a legal document, briefing, compliance summary, or regulatory analysis before it is distributed to clients, stakeholders, or published. Trigger phrases include "verify", "fact-check", "red team", "red-flag", "check accuracy", "validate sources", "quality control", "is this correct", and "review for errors". Produces a structured verification report with severity-categorized errors, verified sources, unsupported claim list, missing disclaimers, and an explicit distribution-readiness assessment.

lawve-ai/awesome-legal-skills Agent Skills author: lawve-ai 3 files body ≈ 4 414 tokens Open the sourcegithub.com↗ analyzed 89 min ago

Adversarial verification of AI-generated legal content with systematic fact-checking, source validation, and quality control.

As a process C 50/100 · Has gaps — weak spots: when it triggers, inputs and preconditions, running it twice

AnalyzerData and analyticsResearchWriting and documentstype and topics are labelled automatically from the skill text
Runs in: Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) OpenClaw Goose OpenCode
JSON
Technical rating
B
94/100
safety, quality, tests
Safety 60%
95
Quality 40%
93
Run on models
none yet
Process rating
C
50/100
Has gaps
Inputs and preconditions w 11
0
Progress reporting w 2
0
When it triggers w 12
20
the three weakest of ten parameters · all ten

How to improve

    For the model run — optional
    • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
    • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

    Guard findings · 5

    ✓ No critical or high findings

    Medium and low: 5
    • low Risky intent intent-offensive-security README.md:1
      Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
      Adversarial verification for AI-generated legal content with systematic fact-checking, source validation, and quality control. Use when verifying legal documents, fact-checking regulatory content, per
    • low Risky intent intent-offensive-security SKILL.md:3
      Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (quoted — discussed, not commanded)
      description: Adversarial verification of AI-generated legal content with systematic fact-checking, source validation, and quality control. Use when a user asks to verify, fact-check, red-team, validat
      quoted
    • low Risky intent intent-offensive-security SKILL.md:10
      Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
      # Red Team Verifier
    • low Risky intent intent-offensive-security SKILL.md:26
      Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
      **Trigger phrases**: verify, fact-check, red team, red-flag, check accuracy, validate sources, quality control, is this correct, review for errors.
    • low Risky intent intent-offensive-security SKILL.md:139
      Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
      # LEGAL RED TEAM VERIFICATION REPORT

    Files scanned: 3. Evidence is masked. Grey chips explain why severity was lowered.

    Against the Agent Skills spec

    ✓ No remarks against the Agent Skills spec

    Process rating: all ten parameters 50/100

    • 0Inputs and preconditions. Does not say what the process needs to start
    • 0Progress reporting. Says nothing while it works
    • 20When it triggers. No condition that starts the skill
    • 30Running it twice. 1 mutating operations with no state check
    • 55Failures and branches. 1 branches
    • 60Tools and files. Uses tools (web) that frontmatter does not declare
    • 60Steps. 86 steps, 4 vague phrases
    • 60Result and completion. Output format stated, no completion criterion
    • 70Execution cost. Instruction body is 4414 tokens
    • 100Consistency. Name and required fields are in place
    • low 15 top-level sections: this looks like several domains in one skill

    Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

    Quality signals

    • +4Description does not say when NOT to use the skill (false activations)
    • +2Single-language instructions
    • +5Description quotes 9 example trigger phrases
    • +3Description length 715: enough signal without eating the budget
    • +4Structure: 44 headings
    • +3Step-by-step instructions: 86 items
    • +3Output format is stated explicitly
    • +4Has examples (1 code blocks)
    • +1License stated

    Quality base 70; lint remarks subtract, signals add up to 100. Result: 93.