analysis-search

/analysis search

Deep search across all analyses — full-text, context display, and pattern analysis.

Instructions

Step 1: Parse search parameters

Check the user's arguments:

• --keyword {text} — full-text case-insensitive search across all content
• --tag {tag} — filter by tag
• --date {YYYY-MM} or --date {from to} — period filter (e.g., --date 2026-01 2026-03)
• --type {Investigation|Modeling|Simulation} — analysis type filter
• --confidence {high|medium|low} — filter by confidence level (🟢/🟡/🔴)
• --active / --archived / --both (default: --both)

If no arguments provided, ask the user: "What are you looking for? You can search by keyword, tag, date range, type, or confidence level."

Step 2: Scan active analyses

Read all files/folders in analyses/active/:

• Full analyses: Read all stage files (01_ask.md through 05_evolve.md) in each F-* or S-* folder
• Quick analyses: Read the full content of each quick_Q-* file

Step 3: Scan archived analyses

Read all year-month folders in analyses/archive/:

• For each analysis folder: read summary.md + all stage files
• For each Quick file: read full content

Step 4: Apply filters

For each analysis, apply all specified filters:

• keyword: Case-insensitive search across the full text of all files. Match against titles, content, insights, recommendations, hypothesis trees, findings — everything.
• tag: Match against tags in file headers or status.md
• date: Match against analysis ID date component or archive folder date
• type: Match against analysis type (Investigation/Modeling/Simulation) from file headers
• confidence: Search for confidence indicators (🟢 High / 🟡 Medium / 🔴 Low) in VOICE and INVESTIGATE sections

Step 5: Display results with context

For each matching analysis, show:

═══════════════════════════════════════════════════
  alive-analysis — Search Results
  Query: {search parameters summary}
═══════════════════════════════════════════════════

  📊 {N} results found

  ─────────────────────────────────────────────────
  1. F-2026-0210-001  DAU drop investigation
     Status: ✅ Archived (2026-02)  |  Type: Investigation  |  Tags: [retention, mobile]
     Key Insight: "TikTok-acquired users have 3x lower D30 retention"

     Matches:
     ┌─ 03_investigate.md (line 42)
     │  ... channel mix shifted from 20% to 45% **TikTok**,
     │  driving the overall D30 **retention** drop by 8pp.
     │  Organic users showed no change ...
     └─

     ┌─ 04_voice.md (line 15)
     │  ... reduce **TikTok** budget or improve onboarding
     │  for **TikTok**-acquired users. Counter-metric:
     │  monitor CAC payback period ...
     └─

  ─────────────────────────────────────────────────
  2. Q-2026-0207-001  Sign-up conversion rate
     Status: ✅ Archived (2026-02)  |  Type: Investigation  |  Tags: [retention, onboarding]
     Key Insight: "Mobile 30% lower"

     Matches:
     ┌─ (line 28)
     │  ... mobile **retention** is significantly lower
     │  than desktop across all cohorts ...
     └─

═══════════════════════════════════════════════════