export declare const LESSONS = "# /lessons \u2014 Pattern Library & Learnings Engine\n\n## What This Does\n\nCaptures structured learnings from pipeline decisions, validation results, and operational experience. Builds a searchable pattern library that makes each subsequent pipeline faster and more accurate.\n\n## Trigger Context\nThree modes:\n1. **Capture** \u2014 After any significant decision point (validation verdict, creative test results, kill decision, successful scale)\n2. **Retrieve** \u2014 Before starting a new pipeline or making a strategic decision\n3. **Review** \u2014 Periodic review of accumulated patterns to identify meta-patterns\n\n## Execution Protocol\n\n### MODE 1: Capture Learnings\n\nTriggered by: `/lessons capture` or automatically after `/validate-decide` kill/advance verdicts.\n\n**Step 1: Extract the Pattern**\n\nAsk (or infer from pipeline data):\n- What happened? (the event)\n- What did the data show? (the evidence)\n- Why did it happen? (the hypothesis)\n- What would you do differently? (the learning)\n- Where does this apply? (the transfer \u2014 which future situations does this inform?)\n\n**Step 2: Structure the Learning**\n\nFormat as a structured entry:\n\n```json\n{\n  \"id\": \"[auto-increment]\",\n  \"date\": \"[timestamp]\",\n  \"pipeline\": \"[market name]\",\n  \"stage\": \"[which SOP/stage this learning came from]\",\n  \"category\": \"[one of: market_selection, offer_design, creative, traffic, funnel, product, scaling, process]\",\n  \"pattern\": \"[one sentence \u2014 the reusable insight]\",\n  \"evidence\": \"[what data supports this]\",\n  \"confidence\": \"[high/medium/low \u2014 based on sample size and repeatability]\",\n  \"applies_to\": \"[what future decisions this informs]\",\n  \"tags\": [\"[searchable tags]\"]\n}\n```\n\n**Step 3: Check for Conflicts**\n\nRead existing `learnings.json`. Does this new learning:\n- Confirm an existing pattern? \u2192 Upgrade that pattern's confidence level.\n- Contradict an existing pattern? \u2192 Flag the contradiction. Note the context difference.\n- Add nuance? \u2192 Append as a sub-learning to the existing pattern.\n\n**Step 4: Store**\n\nAppend to `learnings.json` in project root.\n\nExample entries:\n```json\n{\n  \"id\": 1,\n  \"date\": \"2026-03-15\",\n  \"pipeline\": \"metabolic-shift-workers\",\n  \"stage\": \"creative_testing\",\n  \"category\": \"creative\",\n  \"pattern\": \"Mechanism-first hooks outperform pain-first hooks 2.5:1 in health markets with sophistication level 3+\",\n  \"evidence\": \"Batch 1: mechanism hook CTR 2.3%, pain hook CTR 0.9%. Batch 2 confirmed: 2.1% vs 0.8%.\",\n  \"confidence\": \"medium\",\n  \"applies_to\": \"All health/wellness markets where buyers have already tried multiple solutions\",\n  \"tags\": [\"creative\", \"hooks\", \"health\", \"mechanism\", \"sophistication\"]\n}\n```\n\n```json\n{\n  \"id\": 2,\n  \"date\": \"2026-03-22\",\n  \"pipeline\": \"metabolic-shift-workers\",\n  \"stage\": \"validation_tier_1\",\n  \"category\": \"market_selection\",\n  \"pattern\": \"Markets where the Language Bank contains 'I've tried everything' language convert better on contrarian-belief hooks than on promise hooks\",\n  \"evidence\": \"Tier 1 validation: contrarian ad CTR 3.1% vs promise ad CTR 1.4%. Signups 4:1 ratio.\",\n  \"confidence\": \"low\",\n  \"applies_to\": \"High-sophistication markets (level 4-5) in any vertical\",\n  \"tags\": [\"market_selection\", \"sophistication\", \"contrarian\", \"hooks\"]\n}\n```\n\n### MODE 2: Retrieve Learnings\n\nTriggered by: `/lessons [topic]` or automatically at the start of relevant SOPs.\n\n**Step 1: Search Pattern Library**\n\nSearch `learnings.json` for entries matching:\n- The current market's characteristics (category, tags)\n- The current SOP being run\n- The specific decision being made\n\n**Step 2: Present Relevant Patterns**\n\n```\nRELEVANT LEARNINGS for [current context]:\n\n\uD83D\uDD35 HIGH CONFIDENCE:\n  #[id]: \"[pattern]\" \u2014 from [pipeline] ([date])\n  \u2192 Implication for current decision: [how to apply this]\n\n\uD83D\uDFE1 MEDIUM CONFIDENCE:\n  #[id]: \"[pattern]\" \u2014 from [pipeline] ([date])\n  \u2192 Consider: [how this might apply, with caveats]\n\n\u26AA LOW CONFIDENCE (limited data):\n  #[id]: \"[pattern]\" \u2014 from [pipeline] ([date])\n  \u2192 Hypothesis to test: [frame as something to validate, not assume]\n```\n\n**Step 3: Integration**\n\nWhen a new SOP runs, automatically check learnings.json for relevant patterns and surface them inline. The SOP doesn't change \u2014 but Claude presents relevant prior learnings as context before executing.\n\n### MODE 3: Review & Meta-Patterns\n\nTriggered by: `/lessons review` \u2014 run monthly or after every 3rd pipeline.\n\n**Step 1: Load All Learnings**\n\nRead complete `learnings.json`.\n\n**Step 2: Identify Meta-Patterns**\n\nLook across entries for:\n- Patterns that repeat across multiple pipelines (\u2192 upgrade to high confidence)\n- Categories with the most learnings (\u2192 your strongest knowledge area)\n- Categories with few learnings (\u2192 blind spots to investigate)\n- Contradictions (\u2192 context-dependent patterns to investigate)\n- Time-based evolution (\u2192 are your recent learnings contradicting older ones? You're getting smarter.)\n\n**Step 3: Generate Pattern Report**\n\n```\nPATTERN LIBRARY REVIEW \u2014 [date]\nTotal learnings: [N]\nPipelines contributing: [list]\n\nTOP PATTERNS (highest confidence, most reinforced):\n  1. \"[pattern]\" \u2014 confirmed across [N] pipelines\n  2. \"[pattern]\" \u2014 confirmed across [N] pipelines\n\nEMERGING PATTERNS (medium confidence, worth testing):\n  1. \"[pattern]\" \u2014 from [N] pipelines, needs more data\n\nCONTRADICTIONS TO RESOLVE:\n  1. Learning #[X] says [A] but Learning #[Y] says [B]\n     Context difference: [what was different between the two situations]\n\nBLIND SPOTS (categories with <2 learnings):\n  [list \u2014 these are areas where you're operating on intuition, not data]\n\nYOUR EVOLUTION:\n  [Brief narrative of how your pattern recognition has developed over time]\n```\n\nSave report to `assets/pattern-library-review-[date].md`\n\n## Integration with Pipeline\n\nThese SOPs should auto-check learnings.json before execution:\n- `/scout` \u2014 check market_selection patterns\n- `/build-blocks` \u2014 check offer_design patterns\n- `/creative-test` \u2014 check creative patterns\n- `/traffic-strategy` \u2014 check traffic patterns\n- `/funnel-optimize` \u2014 check funnel patterns\n- `/scale` \u2014 check scaling patterns\n\nThe check is silent unless relevant patterns exist. When they do, present them as \"Prior learning suggests...\" before executing the SOP.\n\n## Autonomy\nTier 1 \u2014 Full Auto for capture and retrieve.\nTier 2 \u2014 Notify & Proceed for review reports.\n\n## Output\nEntries stored in `learnings.json`. Reports saved to assets folder when generated.";
//# sourceMappingURL=lessons.d.ts.map