agentleFS
Sign inStart free
longwave-ai-lablongwave-ai-lab/agentic-workflow-os/CLAUDE.md

agentic-workflow-os

CLAUDE.md0 starsChanged 16 days agoMITView on GitHub

Drop-In Constitutional Rules for Google Antigravity, Cursor IDE, Claude Code & ChatGPT Engineered by Longwave AI Lab (longwave-ai-lab) | Root Constitution (<= 200 Lines Compliant) To prevent context-window bloat and Lost-in-the-Middle attention decay, this root constitution stays strictly under 200 lines. Full specifications, 10-pair execution tables, and ready-to-run Node.js linters (exit code 1) are modularized inside .cursor/rules/*.mdc, .claude/skills/*/SKILL.md, and .agents/skills/*/SKILL.md:

The file

We split this CLAUDE.md into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Multi-Turn 3-Lock Reflex at Recency Anchor (Always Active): Place core execution reflexes at the bottom of root constitutions so short …

    line 20
  2. Very specific:

    Verify package & linter AST: node ./bin/cli.mjs --check

    line 25

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

CLAUDE.md - Agentic Workflow OS (Free Starter Pack (Module 01 - MIT))

Drop-In Constitutional Rules for Google Antigravity, Cursor IDE, Claude Code & ChatGPT Engineered by Longwave AI Lab (longwave-ai-lab) | Root Constitution (<= 200 Lines Compliant)

Quick Routing & Progressive Disclosure Map

To prevent context-window bloat and Lost-in-the-Middle attention decay, this root constitution stays strictly under 200 lines. Full specifications, 10-pair execution tables, and ready-to-run Node.js linters (exit code 1) are modularized inside .cursor/rules/*.mdc, .claude/skills/*/SKILL.md, and .agents/skills/*/SKILL.md:

Module 01: boost-context-bloat-optimizer

  • Summary: Enforces /boost context-window hygiene: <= 200 line root constitutions (CLAUDE.md/GEMINI.md), 3-tier lazy-loaded .mdc and SKILL.md sharding, subagent search offloading, 60% context compaction, and .boost-context-budget.json verification.
  • Cursor Rule: .cursor/rules/01-boost-context-bloat-optimizer.mdc
  • Claude / Antigravity Skill: .claude/skills/boost-context-bloat-optimizer/SKILL.md & .agents/skills/boost-context-bloat-optimizer/SKILL.md
  • Full Spec & Node.js Linter: rules/01-boost-context-bloat-optimizer.md & scripts/lint_boost_context_optimizer.mjs

Core Execution Invariants (Always Active)

  1. /boost Context Hygiene (Module 01): Keep CLAUDE.md and GEMINI.md at or below 200 lines (<= 1,800 tokens) via One-In, Consolidation-Out and Option 0 premise audits. Shard domain rules into path-scoped .mdc globs and on-demand SKILL.md files. Exclude node_modules, dist, .next, and lockfiles via .cursorignore. Offload noisy searches to subagents and compact context at 60% capacity.
  2. 4-Layer Gate & 2x2 Adversarial Matrix (Fail-Closed Verification): Separate Layer 1 deterministic HARD checks (< 2s, AST and closed schema z.strictObject passing on a clean Canary sample) from CONTEXTUAL signals. Route SEMANTIC checks to Layer 2 fail-closed arbitration (PASS, HARD_FAIL, GREY_ZONE_WHITE_FLAG, UNDETERMINED vs SUCCESS, NOT_RUN, TOOL_ERROR). Never fabricate fake probabilities (?? 0.92 or 1.0) when Layer 2 is skipped. Verify 2-layer gates across all 4 cells of the 2x2 Adversarial Matrix.
  3. Multi-Turn 3-Lock Reflex at Recency Anchor (Always Active): Place core execution reflexes at the bottom of root constitutions so short follow-up prompts stay anchored to scope locks (git status -sb), closed schema allowlists (zero Prop-Laundering), and fast < 2s targeted verification gates.

Verification Commands

  • Audit /boost context budget: node scripts/lint_boost_context_optimizer.mjs
  • Verify package & linter AST: node ./bin/cli.mjs --check

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free