Skip to content

AGENTS.md

Latest commit

 

History

History
77 lines (57 loc) · 5.48 KB

AGENTS.md

File metadata and controls

77 lines (57 loc) · 5.48 KB
Raw

Third Brain V8.0 Skills — Codex / Antigravity CLI

This repository contains Agent Skills for Codex CLI & Antigravity CLI environments. Place skills in ~/.agents/skills/ or ~/.gemini/skills/.

V8.0 updates the Obsidian Vault to full OCD Gold-Standard standards, adding an explicit 5-Stage Worker Assembly Pipeline (workflows/worker-flows.md and tools/worker_flow_engine.py):

  1. Worker 1 (Ingest): Immutable source creation (sources/YYYY-MM/) + SHA-256 + block anchors ^....
  2. Worker 2 (Cognitive): Gold-Standard concept cards with Mermaid, paradigm matrices, and evidence bounds.
  3. Worker 3 (GraphWeaver): MOC navigation, Home.md / Central Index updates, and Canvas mapping.
  4. Worker 4 (Governance): YAML validation, link & block ref verification, KPI updates, and test suites.
  5. Worker 5 (Deliverable): Actionable outputs (wiki/outputs/), strategic memos, and daily OKR loop completion.

Skills Summary (20 Skills)

📥 Knowledge Pipeline

  • wiki-ingest — STOW pipeline (Source → Transform → Organize → Write-back) with 13-domain concept placement, 5-category entity placement, block refs, Karpathy understanding gate, and post-ingest lint.
  • knowledge-ops — Multi-layer knowledge management with Markdown-first retrieval, optional vector storage, evidence hierarchy, deduplication, Agent/Wiki flywheel, and knowledge debt queues.
  • wiki-lint — Wiki health check for graph integrity, provenance, link health, taxonomy compliance, clipping lifecycle, understanding, and promotion readiness.

🔄 Daily Loop

  • daily-okr — 7 Key Results daily knowledge compound cycle (Input → Cognition → Wiki → Behavior → Creativity → Output → Feedback).
  • cognitive-compile — 8-section deep learning compile framework (Question → Facts → Concepts → Patterns → Conflicts → Hypotheses → Decision → Action).

🎨 Behavior & Creativity

  • behavior-design — Behavior change system with HAS (Human Agency Scale) framework.
  • creativity-engine — Combinatorial ideation (Bending / Breaking / Blending) + minimum experiments.

🔬 Research & Quality

  • deep-research — STOW-compatible research harness with preflight, source/claim ledgers, activity trace, citations, and wiki-ingest handoff.
  • verify-before-claim — Verification-first quality gate. No completion claims without empirical logs or test receipts.

🔄 Learning & Workflow

  • session-learn — Knowledge extraction with Closure Protocol.
  • project-flow-ops — Project triage, state tracking, and execution governance.

📊 Context & Cost

  • context-manager — Runtime-derived context budgets, checkpoints, compaction, retrieval, and capability routing.
  • token-cost-tracker is a command under commands/, not an Agent Skill.

🏗️ Engineering & Multi-Agent Architecture

  • agentic-engineering — Agent-as-process workflow refactoring with autonomy defaults, delegated-action boundaries, state checkpoints, write-back, and verification gates.
  • loop-engineering — Temporal-depth control through bounded Trigger → Execute → Verify → State loops, durable contracts, hard budgets, and stop/recovery rules.
  • graph-engineering — Dependency-width control for bounded static DAGs with explicit dependencies, independent branches, typed joins, and node-local recovery.
  • harness-engineering — Runtime kernel design: scheduler, permissions, tools as system calls, MCP/Skills/Hooks selection, provenance ledgers, observability, and recovery.
  • agent-teams-command — Multi-agent process ownership, IPC, worktree isolation, async budget envelopes, parallel subagent orchestration, and evidence gates.

💼 Strategy & Operations

  • startup-evaluation — Startup health diagnosis with entrepreneurship, VC 5T, PMF, runway, team, and next-test frameworks.
  • anthropic-os — Self-evolving work method engine with 3B creativity algorithms.
  • ai-six-sigma-property-os — AI + Ontology + DMAIC Black Belt operating model for property work orders, dispatch, quotes, evidence, CTQ dashboards, and MVP quality control.

V7.2 Vault Taxonomy Contract (system/config.md)

  • wiki/concepts/ (13 Domains): ai-engineering, ai-economics, ai-science, behavioral-econ, business-strategy, entrepreneurship, general-concepts, geopolitics-energy, identity-culture, investing-macro, investing-quant, investing-vc, knowledge-systems.
  • wiki/entities/ (5 Categories): people, companies, funds-investors, products, orgs.
  • sources/ (6 Pools): YYYY-MM/ (e.g. 2026-07/), pre-2026/, books/ (>100KB).
  • wiki/outputs/ (3 Categories): gmail-digests/, evaluations/, compilations/.
  • maps/ (4 Tiers): domain-mocs/, system-indexes/, project-maps/, canvases/ (plus Home.md & 中央索引.md).

Skill Contract & Execution

When selecting a skill, read its frontmatter before executing:

  • assumes — required operating assumptions.
  • conflicts_with — boundaries that must not be silently overridden.
  • metadata.profile — one-shot, stateful, loop, or high-risk controls.
  • ## Failure Protocol and ## Output Contract — standard stop status and receipt.
  • ## Success Metrics — minimum observable result for one successful run.
  • ## Quality Gates — checks that must pass before claiming completion.

Karpathy LLM OS

LLM=CPU · Context=RAM · Storage=Disk · Tools=System Calls · Skills=Programs · Harness=Kernel · Agent Teams=Processes