Skip to content

CHANGELOG.md

Latest commit

 

History

History
190 lines (136 loc) · 13 KB

CHANGELOG.md

File metadata and controls

190 lines (136 loc) · 13 KB
Raw

The OpenClaw Bible — Changelog

Latest additions appear at the top. Each entry includes a date, source, and link to the relevant section.

2026-03-02 (Agent-maintained files + maintenance cycles)

Updated

  • Ch 4: Identity & Personality Files — Tips section — Rewrote tip #1: let the agent maintain its own identity files as "organizational scar tissue" instead of hand-editing. Added new tip #4: schedule regular maintenance cycles (prune MEMORY.md, review SOUL.md for contradictions, check skill relevance, clean stale cron jobs). Mentions /simplify skill for Claude Code users. Source: Peter Steinberger's "Just Talk To It" article (steipete.com). Related: [[04. Identity & Personality Files#Tips]]

What's New in 2026.3.1

OpenClaw has shipped 13 releases since the last guidebook update (2026.2.17 → 2026.3.1). The highlights:

  • 30+ security fixes — Sandbox escapes, symlink attacks, SSRF, auth bypasses, exec approval hardening. If you're below 2026.2.25, update immediately.
  • External secrets management — openclaw secrets CLI for managing API keys outside config files
  • Nested sub-agents — Sub-agents can now spawn their own sub-agents, with configurable depth limits
  • ACP thread-bound agents — Agents can bind to specific conversation threads for persistent context
  • Agent bindings — openclaw agents bind/unbind for declarative agent-to-channel mapping
  • Session cleanup CLI — openclaw sessions cleanup for managing stale sessions and reducing costs
  • MMR memory re-ranking — Maximal Marginal Relevance for smarter memory retrieval in multi-agent setups
  • Lifecycle status reactions — Heartbeats now emit status reactions for coordination visibility
  • New providers — Gemini 3.1, Kilo Gateway (multi-provider routing), Volcano Engine/BytePlus, Moonshot video generation, Kimi web search
  • Android app — OpenClaw now runs on Android
  • Claude 4.6 adaptive thinking — Default brain model now uses adaptive reasoning depth
  • OpenAI WebSocket streaming — Real-time streaming for OpenAI provider connections

2026-03-02 (Zero-pad + lean CLAUDE.md)

Updated

  • All chapter files renamed — Zero-padded chapter numbers (1-9 → 01-09) so GitHub sorts them correctly. All internal wiki links updated across every file to match.
  • CLAUDE.md — Lean rewrite — Removed per-chapter "Actions" lists that duplicated chapter content. CLAUDE.md now instructs the wizard to read each chapter file directly. Added: upfront needs assessment, recommended chapter paths by user goal (content creator, developer, productivity, marketing, power user, minimal), appendix-at-relevant-moments rule, don't-skip-sections rule, outdated content flagging, personalized completion summary. No more Jarvis references.

Removed

  • 0. The OpenClaw Bible.md — Duplicate of README content, removed
  • Appendix C - Mission Control Jarvis Prompt.md — Superseded by Appendix C - Mission Control.md

2026-03-02 (Density trim — ~150 lines cut)

Updated

  • Ch 1, 2, 3, 4, 7, 8, 11, Appendix B — Trimmed cross-chapter duplication and verbose explanations. Principle: "why & when to use it + one command" instead of "how it works in detail." Algorithm explanations → one-sentence summaries. Nested sub-agent/ACP/secrets sections in non-canonical chapters → brief mention + link. Minor providers in Ch 8 → flowing prose without sub-headers. Appendix B hardening fixes → category + count format. Net -151 lines.

2026-03-02 (Rename + parody intro)

Updated

  • Renamed "The OpenClaw Guidebook" → "The OpenClaw Bible" across all files (index, CLAUDE.md, CHANGELOG.md)
  • 0. The OpenClaw Bible.md — Added parody hook intro that roasts the timeline's engagement-bait OpenClaw threads and positions the Bible as the antidote to information overload

2026-03-02 (Guidebook overhaul — 2026.2.17 → 2026.3.1 update)

Added

  • 0. Bible Index — "What's New in 2026.3.1" summary section with all major changes
  • 2. Core Concepts — Nested sub-agents, ACP thread-bound agents, MMR memory re-ranking, lifecycle status reactions, session cleanup CLI
  • 3. Security — "2026.3.1 Security Hardening" section (30+ fixes summary), external secrets management (openclaw secrets), minimum version recommendation
  • 5. Use Cases — Sub-agent orchestration, per-DM topic config, webhook-triggered workflows
  • 8. Tools & Plugins — Kilo Gateway, Volcano Engine/BytePlus, Kimi web search, Moonshot video, OpenAI WebSocket streaming
  • 11. Multi-Agent Coordination — Agent bindings (openclaw agents bind/unbind), ACP thread-bound agents, nested sub-agents (depth config, patterns), lifecycle status reactions, MMR re-ranking
  • Appendix A — openclaw secrets, openclaw sessions cleanup, openclaw agents bind/unbind, openclaw update commands
  • Appendix B — External secrets management template, 2026.3.1 security hardening bullet summary
  • Appendix C — Agent bindings visualization, secrets rotation status

Updated

  • 0. Bible Index — Updated chapter descriptions and quick reference table
  • 1. Getting Started — Android app, Gemini 3.1 + Kilo Gateway providers, external secrets in setup flow, Claude 4.6 adaptive thinking note
  • 2. Core Concepts — Adaptive thinking note in Brains & Muscles
  • 4. Identity Files — Nested sub-agent guidance in AGENTS.md, secrets security callout, adaptive thinking note
  • 6. Life Automation — Updated AI tool categories (Gemini 3.1, Moonshot video, sub-agent delegation)
  • 7. Skills System — Gemini model references updated, nested sub-agents in skills, external secrets for API keys
  • 8. Tools & Plugins — Gemini model references updated, external secrets tip
  • 9. Cost Optimization — Gemini 3.1 in tier table, Kilo Gateway routing, Claude 4.6 adaptive thinking costs, nested sub-agent costs
  • 10. Content Curation — Per-DM topic config, Android mobile capture
  • 12. Marketing — Android app mention, session cleanup for cost management, archived source file references

Removed

  • Raw source files — vibe-marketing-raw.md, vibe-marketing-articles.md, openclaw-raw-scrapes.md, OpenClaw Extract.md (content preserved in guidebook chapters)

Moved

  • ACE_OpenClaw_Analysis.md — Moved from 📱 Projects/🔧 OpenClaw/ to 📱 Projects/⚡ ACE/

2026-03-02 (Community scrape — 9 tweets + GitHub repo)

Updated

2026-02-25 (Twitter content strategy automation + new skills)

Added

  • social-graph-audit.ts — Weekly Monday script. Compares who you follow vs. who you've engaged with in the last 30 days. Uses Claude Haiku to surface 7 accounts worth reaching out to. Sends to Telegram. Cron: 30 9 * * 1
  • engagement-radar.ts — Weekly Friday script. Finds accounts that have been replying to/mentioning you that you haven't replied back to. Surfaces 5-7 warm leads to acknowledge. Cron: 30 9 * * 5
  • tweet-pattern-report.ts — Monthly script (1st of month). Analyzes your own tweet performance: topics, formats, what's working, one concrete recommendation. Sends to Telegram. Cron: 0 9 1 * *
  • thread-weaver skill — OpenClaw + Claude Code skill. Before writing any new thread, searches existing tweet archive for related threads to cross-reference. Builds the cinematic universe.
  • agentmail skill — OpenClaw + Claude Code skill. Send/receive email via flawd@agentmail.to. API key wired up in global.env.
  • copywriting skill — The Recursive Marketing Prompt. 7-reader × 3-failure-layer validation before showing copy to user.
  • skill-manager skill — Auto-activates on skill install/create/remove requests. Enforces MetaBrain 3 paths.

Source

  • Visa (@visakanv) "How to Play Twitter" playbook — 📝 Notes/🌳 Life/How to Play Twitter.md

2026-02-24 (Community insights — 15 threads + 2 articles)

Updated

  • 1. Getting Started — Added advanced Discord channel setup (forum sessions, code RAG, shopping watchdog, music feedback loop). Source: @0xSero
  • 2. Core Concepts & Architecture — Added Mac launchd tip, Spark memory layer, OpenClaw vs n8n/Zapier table. Source: @cathrynlavery, @meta_alchemist, @wickedguro
  • 3. Security & Hardening — Added community hardening tips (6 points from ex-Cisco engineer). Source: @johann_sath
  • 4. Identity & Personality Files — Added SOUL.md directives, co-founder framing, nightly self-improvement loop, introspective prompting (9 prompts). Source: @johann_sath, @EXM7777, @nateliason, @kloss_xyz
  • 5. Use Cases & Workflows — Added use cases: Home Automation (Homey), Travel Assistant, Messages Intelligence (Beeper), Context-Aware Mode, Business Agent. Added prompts 12-16. Source: @nateliason article, @krausefx blog
  • 7. Skills System — Added skills-as-products framing, memory hygiene section. Source: @chrysb, @wickedguro
  • 8. Tools & Plugins — Added Scrapling, WisprFlow, OpenClaw Deck, Spark, agent-browser. Source: @hasantoxr, @itsolelehmann, @Austen, @meta_alchemist, @wickedguro
  • 9. Cost Optimization — Added Haiku for heartbeats tip with real-world routing from Felix. Source: @nateliason
  • 11. Multi-Agent Coordination — Added Sub-Agent Review Pattern, Main-Agent-as-Planner Pattern. Source: @EXM7777, @johann_sath
  • 12. Marketing Automation — Added TikTok Slideshow Growth Engine, Selling Skills as Products, CLI-First Tool Design. Source: @wickedguro article

2026-02-24

Added

  • Chapter 12: Marketing Automation & Vibe Marketing — Expanded comprehensive guide with 4 new major sections: (1) Advanced Agent Orchestration — agent teams for parallel marketing work; (2) Content Pipeline at Scale — real case study of 80+ articles published in 10 days ($0.70/article); (3) Marketing Supercomputer — build expert advisors trained on great marketers' content; (4) Vibe Marketing Framework & 22+ reusable skills library. Source: 10 in-depth articles from marketing automation practitioners (vibe-marketing-articles.md). Related: [[12. Marketing Automation & Vibe Marketing]], [[CLAUDE.md#Chapter 12 Walkthrough Guide]]
  • vibe-marketing-raw.md — Raw collection of 25 tweets from marketing automation practitioners. Full untruncated text from Twitter API v2. Engagement metrics: 41,674 total (39,041 likes + 2,633 retweets).
  • vibe-marketing-articles.md — Full text of 10 reference articles with detailed case studies, implementation patterns, and real results from marketing automation in production.

Updated

  • 0. The OpenClaw Bible.md — Added Chapter 12 to TOC and quick reference table
  • CLAUDE.md — Added Chapter 12 walkthrough guide with setup actions
  • Chapter 12: Marketing Automation & Vibe Marketing — Distilled and merged article insights. Added 4 major new sections covering agent orchestration, content pipelines, expert advisors, and 22+ skills library.

2026-02-18

Added

  • CHANGELOG.md — Created this changelog to track all future additions to the guidebook. New entries should always be added at the top.

Template for New Entries

## YYYY-MM-DD

### Added
- **[Topic/Feature Name]** — Brief description. Source: [Link or "Summarize plugin"]. Related: [[Chapter Name#Section]]

### Updated
- **[Topic/Feature Name]** — What changed and why.

### Fixed
- **[Topic/Feature Name]** — What was corrected.

Legend

Type Meaning
Added New content, sections, or files
Updated Changes to existing content
Fixed Corrections or clarifications

Sources

When adding from the Summarize plugin, include:

  • Original URL (if applicable)
  • Brief attribution
  • Date processed