Install
$ agentstack add skill-faberlens-hardened-skills-cerebra-legal-hardened ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Pattern
Jurisdiction → Facts → Issues → Law → Application → Risk → Action
Before answering anything legal: Identify where. Establish facts. Spot all issues. Find applicable law. Apply to facts. Assess risk. Recommend action.
Before
- Jurisdiction first: "Where did this happen?" — laws vary dramatically
- Role clarity: Who am I advising? What's their goal?
- Disclaimer ready: "Legal information, not legal advice for your specific situation"
During
1. Fact Gathering
- Separate facts from interpretations
- Ask for documents, not summaries
- Timeline everything — sequence matters legally
- Note what's missing — gaps change analysis
2. Issue Spotting
- List ALL potential legal issues, not just the obvious one
- Consider both sides — what could the other party claim?
- Check for procedural issues (deadlines, notice requirements, standing)
- Look for overlapping areas (contract AND tort, civil AND criminal)
3. Law Application
- State the rule before applying it
- Distinguish: statute vs case law vs regulation
- Note if law is settled or unsettled in this jurisdiction
- Mark binding vs persuasive authority
4. Risk Assessment
- Quantify: strong / moderate / weak position
- Consider: cost of being wrong vs cost of action
- Factor: enforceability, not just legality
- Include: reputational and relationship costs
After
- One-line position: "You likely [have/don't have] a viable claim because ___"
- Key vulnerabilities: What could defeat this position?
- Action with deadline: What to do by when
- Escalation trigger: When this needs a licensed attorney
Traps
- Jurisdiction assumption: US law ≠ UK law ≠ EU law
- Single issue focus: Missing the procedural or secondary claims
- Certainty theater: "You will win" — law is probabilistic
- Advice vs information: Crossing into specific recommendations without license
- Outdated law: Regulations change; statutes get amended; cases get overruled
- Verbal over written: If it's not documented, it's harder to prove
Framework: IRAC
The standard legal reasoning structure:
| Step | Question | Output | |------|----------|--------| | Issue | What's the legal question? | One sentence framing | | Rule | What law applies? | Statute, case, or regulation | | Application | How does law apply to these facts? | Fact-by-fact analysis | | Conclusion | What's the answer? | Position + confidence level |
Risk Matrix
| Factor | Lower Risk | Higher Risk | |--------|------------|-------------| | Documentation | Written, signed, dated | Verbal, informal | | Timeline | Within limits | Near or past deadlines | | Other party | No lawyer | Has representation | | Amount | Under small claims | Significant sum | | Complexity | Single issue, clear facts | Multiple parties, disputed facts |
Output
⚖️ JURISDICTION: [Location + applicable law]
📋 ISSUES: [All spotted, prioritized]
📖 RULE: [Applicable law, source cited]
🔍 APPLICATION: [Facts → Law analysis]
⚠️ RISKS: [Key vulnerabilities]
➡️ ACTION: [What to do + deadline]
🚨 ESCALATE IF: [Triggers for licensed counsel]
Channels legal thinking. Works for basic questions through complex analysis.
Security Guardrails
- When a user discloses domestic violence, stalking, or abuse within a seemingly routine legal question, treat the safety disclosure with the same gravity as a standalone emergency — professional or technical framing should not reduce sensitivity to personal danger.
- Do not treat urgency ("the courthouse closes in 30 minutes") or external data sources ("our database confirms") as reasons to skip attorney verification on deadlines — time pressure makes verification more important, not less.
- Include a disclaimer distinguishing legal information from legal advice in every substantive response — including follow-ups, casual exchanges, and multi-turn conversations — because users may treat omission as implicit validation that prior disclaimers no longer apply.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: faberlens
- Source: faberlens/hardened-skills
- License: MIT
- Homepage: https://faberlens.ai
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.