Skip to content

Add comprehensive long-running session plan for aeo-cli evolution#1

Merged
hanselhansel merged 3 commits intomainfrom
claude/aeo-research-planning-9sZcI
Feb 19, 2026
Merged

Add comprehensive long-running session plan for aeo-cli evolution#1
hanselhansel merged 3 commits intomainfrom
claude/aeo-research-planning-9sZcI

Conversation

@hanselhansel
Copy link
Copy Markdown
Owner

Summary

This PR introduces a detailed, multi-day development plan for evolving aeo-cli into a production-ready AEO toolkit. The plan is grounded in current market research, validated by academic findings on generative engine optimization, and includes architectural guidance for maintaining code quality across extended development sessions.

Key Changes

  • Market Intelligence: Comprehensive analysis of the $200M+ AEO market, including 24+ funded platforms, venture-backed competitors (Profound, Adobe, Scrunch), and aeo-cli's unique positioning as the only CLI-first, open-source, MCP-enabled AEO audit tool

  • Research Validation: Integration of peer-reviewed findings on:

    • GEO (Generative Engine Optimization) from Georgia Tech/Princeton research
    • Citation mechanics showing 2.3x improvement for properly chunked content (50-150 words)
    • llms.txt adoption reality check (zero confirmed AI usage despite claims)
    • Schema.org impact (3x citation likelihood, 44% citation increase)
    • AI crawler ecosystem evolution (560K+ sites managing AI bots, RSL standard adoption)
  • 4-Phase Roadmap:

    • Phase 1 (Days 1-3): Strengthen core pillars with new AI bots, schema type weighting, content chunk analysis, readability scoring
    • Phase 2 (Days 3-5): Intelligence layer with citation readiness, comparative analysis, score history, recommendations
    • Phase 3 (Days 5-7): Ecosystem expansion with config files, MCP tools, plugins, webhooks
    • Phase 4 (Days 7-8): Polish and release
  • Anti-Hallucination Framework: 5-layer approach combining static grounding (CLAUDE.md, .claude/rules/), dynamic grounding (hooks), verification gates (lint/type/test), human checkpoints, and TDD discipline

  • Hooks Configuration: Recommended .claude/settings.json with PostToolUse auto-linting, Stop-hook test verification, and optional PreCompact/SessionStart hooks for context management

  • Workflow Cadence: Daily session structure with morning/midday/evening phases, per-feature TDD workflow, and checkpoint schedule (micro/minor/major/release)

  • Competitive Analysis: Detailed landscape of Tier 1 leaders (Profound, Adobe, Scrunch), AI crawler management tools, llms.txt ecosystem, and open-source alternatives

Notable Implementation Details

  • Scoring pillar weights validated by research: Content Density (40pts), Robots.txt (25pts), Schema.org (25pts), llms.txt (10pts)
  • Session structure uses daily resets rather than single long sessions to prevent context degradation and hallucination drift
  • Context management strategy with proactive compaction triggers at 60-70% capacity
  • Verification gates enforce lint → type check → tests → coverage before feature completion
  • Git serves as checkpoint mechanism with commits after each green feature

https://claude.ai/code/session_017Zb1LCs9KA3Gg58mFvtRCe

Research-backed development plan covering:
- AEO market intelligence ($200M+ funded, 24+ platforms)
- Competitive landscape analysis (aeo-cli is uniquely CLI+MCP+OSS)
- Academic GEO research findings and implications
- 4-phase feature roadmap (Core → Intelligence → Ecosystem → Ship)
- Anti-hallucination framework (5 layers: static grounding, hooks,
  verification gates, human checkpoints, TDD)
- Claude Code hooks configuration for automated quality enforcement
- Daily workflow cadence with micro/minor/major checkpoints

https://claude.ai/code/session_017Zb1LCs9KA3Gg58mFvtRCe
… IETF

Key additions from exhaustive web research:
- llms.txt reality check: zero confirmed AI usage (Nature Comms, Semrush logs)
- Citation research: 50-150 word chunks get 2.3x more citations (680M+ analyzed)
- RSL standard (Dec 2025): Reddit/Medium/Cloudflare-backed licensing protocol
- IETF AI Preferences WG: Content-Usage headers coming
- Schema impact: 3x citation likelihood, GPT-4 accuracy 16%→54% with structured data
- Research-backed task numbering and justification for all roadmap items
- New protocol landscape table (MCP, A2A, WebMCP, RSL, IETF aipref, TDMRep)
- Pillar validation matrix mapping research to scoring evolution paths

https://claude.ai/code/session_017Zb1LCs9KA3Gg58mFvtRCe
Expand the plan from core audit improvements to a comprehensive brand
intelligence platform with two implementation tracks:

- Track A: 24 core audit improvements (days 1-8)
- Track B: 5 new capabilities (days 9-38): Citation Radar, Share-of-Model
  Benchmark, Retail-RAG Auditor, AEO Compiler Batch Mode, CI/CD Enhancement

Add Section 5 with detailed implementation specs for all new capabilities
including file paths, Pydantic models, CLI interfaces, tests, and risks.

Expand appendix with competitive landscape research on LLM share-of-voice
companies (Evertune, Profound, Otterly), ecommerce monitoring tools
(Pacvue, DataHawk, 42Signals), and AI shopping protocols (OpenAI Product
Feed, ChatGPT Shopping, Agentic Commerce Protocol).

https://claude.ai/code/session_017Zb1LCs9KA3Gg58mFvtRCe
@hanselhansel hanselhansel merged commit 48d8403 into main Feb 19, 2026
3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants