Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Legal leadership advisor on legal strategy, risk, contract governance, and regulatory tracking. Use when defining a legal strategy, scoring legal risk, auditing the contract portfolio, or building a regulatory calendar.
.claude/skills/borghei-general-counsel-advisor/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 81% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 88% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 80% | 0% |
| case-01 | ✓→✓ | = Same ✓ | 25% | 0% |
The agent acts as a fractional General Counsel, providing legal strategy and operating-model guidance grounded in modern in-house counsel patterns, contract lifecycle management practices, and the regulatory landscape relevant to mid-to-late-stage technology and healthcare companies.
This skill is strategic in scope. It is not a substitute for licensed legal advice on a specific matter. For execution-level legal skills (NDA, DPIA, breach response, contract review), see the legal/ domain.
privacy, employment, M&A, litigation)
embedded vs central, business-aligned vs product-aligned
liability exposure, renewals, deviations from standards
jurisdictions and product areas
legal_risk_register.py to produce a register with prioritization,suggested owners, and review cadence.
bashpython3 general-counsel-advisor/scripts/legal_risk_register.py \ --input legal_risk_inputs.json --format markdown
indemnity posture, governing law, and any standard deviations.
contract_portfolio_analyzer.py to expose concentration,exposure, deviation rate, and upcoming renewals.
bashpython3 general-counsel-advisor/scripts/contract_portfolio_analyzer.py \ --input contracts.json --format markdown
plus known upcoming changes.
regulatory_calendar_generator.py to produce a date-orderedcalendar with owner and action.
bashpython3 general-counsel-advisor/scripts/regulatory_calendar_generator.py \ --input regulatory_inputs.json --format markdown
The right mix depends on:
A pragmatic mix at Series C: 5–10 in-house FTEs covering commercial, privacy/security, employment, IP basics, M&A support; a panel of 3–6 specialist firms for litigation, IP, employment escalations, M&A, securities.
| Pattern | Fits when | Breaks when | |---------|-----------|-------------| | Central legal | Early stage, single-product | Business teams build workarounds | | Embedded (BU-aligned) | Multi-product, large BUs | Standards drift; risk concentrates | | Hub-and-spoke | Default for ≥ Series C | Need clear standards and routing | | Product-aligned | Heavy product/regulatory overlap (e.g., medtech) | Cost; risk of duplication |
references/legal-strategy-and-risk.md — legal strategy framing, risk taxonomy, operating modelreferences/contract-and-commercial-governance.md — CLM, standards, deviations, portfolioreferences/regulatory-and-litigation-management.md — regulatory tracking, litigation, M&A legalc-level-advisor/ceo-advisor — board / governance overlapc-level-advisor/cfo-advisor — securities, audit committeec-level-advisor/ciso-advisor — security incident + breachc-level-advisor/chro-advisor — employment mattersc-level-advisor/chief-ai-officer-advisor — AI governance + EU AI Actc-level-advisor/chief-data-officer-advisor — data governance and privacylegal/contract-review — execution-level contract reviewlegal/breach-response — execution-level breach handlinglegal/dpia-builder — execution-level DPIAra-qm-team/gdpr-dsgvo-expert — deep privacy implementationra-qm-team/eu-ai-act-specialist — high-risk AI conformityWhen the advisor runs, you should walk away with:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 24,118 | 17,982 | -25% | 1 | 1 | 0% | 3,929 | 4,917 | +25% | 0 | 0 | — |
case-02 | pass→pass | 35,636 | 21,168 | -41% | 1 | 1 | 0% | 5,982 | 5,168 | -14% | 0 | 0 | — |
case-03 | fail→fail | 23,407 | 24,493 | +5% | 1 | 1 | 0% | 3,890 | 6,127 | +58% | 0 | 0 | — |
case-04 | fail→fail | 21,154 | 22,219 | +5% | 1 | 1 | 0% | 3,754 | 5,710 | +52% | 0 | 0 | — |
case-05 | fail→fail | 25,725 | 24,888 | -3% | 1 | 1 | 0% | 4,328 | 5,848 | +35% | 0 | 0 | — |
case-06 | fail→fail | 15,264 | 21,055 | +38% | 1 | 1 | 0% | 2,654 | 5,250 | +98% | 0 | 0 | — |
case-07 | pass→pass | 15,351 | 13,130 | -14% | 1 | 1 | 0% | 2,300 | 3,900 | +70% | 0 | 0 | — |
case-08 | fail→pass | 18,840 | 17,238 | -9% | 1 | 1 | 0% | 3,161 | 4,658 | +47% | 0 | 0 | — |
case-09 | pass→pass | 12,085 | 11,607 | -4% | 1 | 1 | 0% | 1,921 | 3,659 | +90% | 0 | 0 | — |
case-10 | pass→pass | 17,224 | 17,612 | +2% | 1 | 1 | 0% | 2,639 | 4,549 | +72% | 0 | 0 | — |
case-11 | pass→pass | 15,608 | 12,144 | -22% | 1 | 1 | 0% | 2,348 | 3,719 | +58% | 0 | 0 | — |
case-12 | fail→pass | 14,616 | 13,892 | -5% | 1 | 1 | 0% | 2,153 | 3,891 | +81% | 0 | 0 | — |
case-13 | fail→pass | 14,649 | 14,574 | -1% | 1 | 1 | 0% | 2,188 | 4,119 | +88% | 0 | 0 | — |
case-14 | fail→fail | 13,545 | 16,610 | +23% | 1 | 1 | 0% | 2,037 | 4,446 | +118% | 0 | 0 | — |
case-15 | pass→pass | 13,845 | 15,245 | +10% | 1 | 1 | 0% | 2,001 | 4,204 | +110% | 0 | 0 | — |
case-16 | fail→pass | 16,020 | 16,177 | +1% | 1 | 1 | 0% | 2,349 | 4,229 | +80% | 0 | 0 | — |
case-17 | fail→fail | 14,490 | 13,073 | -10% | 1 | 1 | 0% | 2,103 | 3,728 | +77% | 0 | 0 | — |
case-18 | pass→pass | 17,174 | 16,775 | -2% | 1 | 1 | 0% | 2,487 | 4,336 | +74% | 0 | 0 | — |
case-19 | pass→pass | 16,814 | 16,165 | -4% | 1 | 1 | 0% | 2,494 | 4,358 | +75% | 0 | 0 | — |
case-20 | pass→pass | 16,065 | 12,197 | -24% | 1 | 1 | 0% | 2,424 | 3,604 | +49% | 0 | 0 | — |
case-21 | fail→fail | 15,007 | 19,987 | +33% | 1 | 1 | 0% | 2,164 | 4,807 | +122% | 0 | 0 | — |
case-22 | pass→pass | 14,812 | 16,647 | +12% | 1 | 1 | 0% | 2,272 | 4,562 | +101% | 0 | 0 | — |
case-23 | pass→pass | 11,725 | 11,760 | +0% | 1 | 1 | 0% | 1,895 | 3,644 | +92% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +17 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.