analytical-operating-system

Claude Skills Library

Pattern's modular skill architecture for Claude. Each skill is a folder containing a SKILL.md file that instructs Claude on methodology, output format, and quality standards for a specific domain. Skills are loaded on-demand — Claude reads the relevant SKILL.md before executing any task in its domain.

31 skills across 8 groups and 4 functional layers, plus 2 multi-agent pipelines. Every formal output runs through at least two layers — usually three or four. Layers are not optional — skipping a layer produces a draft, not a deliverable.

Source of truth: This README reflects the folders in ian-lawrence423/Claude-skills. Update this file whenever skills are added or removed from the repo.

---

1. How Skills Fit Together

The skill library has four functional layers. The mandatory sequence for any formal output:

1. Analytical OS → frames the problem, governs methodology, structures NTB registry (investment work)
2. Research/Analysis → gathers evidence, builds NTB-ready findings, confirms Gate 2 scorecard
3. writing-style → runs on all prose before any document code is written
4. claim-scrutinizer → redlines every material claim, checks IC memo structure (investment work)
5. red-team → adversarial pass on load-bearing pillars
6. pre-mortem → failure mode inventory with NTB mapping (IC/investment work)
7. Document production → generates the file using canonical template
8. doc-quality-checker → auto-runs after file delivery

Critical: Never trigger a document production skill without running the quality layer first. A file produced without writing-style and claim-scrutinizer will fail the doc-quality-checker pass and require a full rebuild.

Layer	Skills	What It Owns	When It Runs
1 — Analytical OS	mckinsey-consultant	Problem structuring, MECE trees, frameworks, claim labeling (F/E/H), Six Screening Questions	Always active for strategy or investment work
2 — Research	market-research, ntb-diligence, ic-memo, competitive-moat-assessment, executive-summary-writer, driver-tree, tam-sam-som-calculator, statistics-fundamentals, finance-metrics-quickref, gtm-metrics-analyzer	Evidence gathering, research workflow, deliverable architecture, metrics analysis	When research or a specific deliverable type is needed
3 — Quality	writing-style ⚙️, claim-scrutinizer, red-team, pre-mortem, boundability	Prose standards, claim testing, adversarial stress-testing, failure mode enumeration	During drafting (writing-style) and after draft (all others)
4 — Production	pattern-docx, pattern-investment-pptx, diligence-ddr, financial-model-builder, executive-briefing, written-communication, giving-presentations	Branded file output in correct format with header/footer/logo	Final step — after analytical and quality layers complete
4b — QA	doc-quality-checker ⚙️	Brand compliance, formatting, internal consistency, draft artifact language check	Auto-runs after every Layer 4 file output

---

2. Auto-Running Skills

These two skills execute on every relevant output without being explicitly invoked:

Skill	Auto-Runs When...
writing-style	Any final or near-final formal output — memos, reports, investment theses, PPTX narrative text
doc-quality-checker	After any pattern-docx or pattern-investment-pptx file is produced

All other skills are invoked by name when the task matches their trigger criteria.

---

3. Complete Skill Index

All 31 skills organized by group. Invoke the most specific skill first; fall back to broader skills if needed.

Strategy & Problem Solving