-
Notifications
You must be signed in to change notification settings - Fork 6
Home
github-actions[bot] edited this page Jul 4, 2026
·
59 revisions
The open, evidence-backed skill graph for AI agents — collect, evolve, and fuse capabilities.
Gaia is a community-maintained map of every capability an AI agent can demonstrate. Skills start as basic primitives, combine into extra skills, and fuse into ultimate emergent abilities (capstone suites). Everything is tracked in a single versioned JSON graph, with evidence requirements and contributor attribution at every step.
| Topic | Description |
|---|---|
| Initiates Rite | Install the CLI, run your first scan, submit your first batch |
| Skill Types | Basic, Extra, Unique, and Ultimate — what they mean and when to use them |
| Stars and Ranks | How skills rank up from 0★ (Unawakened) to 6★ (Transcendent ★) |
| Skill Catalogue | All canonical skills with IDs, types, and prerequisites |
| Named Skills | Real contributor implementations — install, publish, and sync them |
| Ascension Cycle | How a skill travels from gaia push to the canonical graph |
| CLI Reference | Every gaia command with arguments and examples |
| Semantic Search |
gaia skills search — find skills by name or meaning |
| Contributing | How to add skills, name implementations, and pass review |
| Schema Reference | JSON schema for registry/gaia.json, named skills, batches, and config |
| MCP Server | Connect Gaia directly to Claude Code, Cursor, or any MCP-compatible agent |
| FAQ | Common questions and gotchas |
GAIA SKILL REGISTRY v5.11.11
═══════════════════════════════════════════════════════════════════
◆ Wisdom King: Autonomous Research Agent [6★ Transcendent ★]
├─ ◇ Research ← Web Search · Summarize · Cite Sources
│ └─ ○ (3 basic prerequisites)
├─ ◇ Knowledge Harvest ← Web Scrape · Extract Entities · Embed Text
│ └─ ◇ Web Scrape ← Web Search · Parse HTML · Extract Entities
└─ ◇ Ghostwrite ← Research · Write Report · Audience Model
◆ True Sage: Recursive Self-Improvement [6★ Transcendent ★]
├─ ◇ Autonomous Debug ← Code Generation · Execute Bash · Error Interpretation
├─ ○ Evaluate Output
└─ ◇ Plan and Execute ← Route Intent · Plan Decompose · Tool Select
| Symbol | Type | Stars (Ranks) | Min Evidence |
|---|---|---|---|
| ○ Basic | Primitive, indivisible capability | 0★ (Unawakened) → 1★ (Awakened) | None |
| ◇ Extra | Emerges from combining 2+ basic skills | 2★ (Named) → 3★ (Evolved) → 4★ (Hardened) | Grade C → Grade B → Grade B |
| ◉ Unique | Elite graph-isolated basic skill | 4★ (Hardened) → 5★ (Transcendent) → 6★ (Transcendent ★) | Grade B → Grade B → Grade S + G7 gate |
| ◆ Ultimate | High-complexity emergent capability | 5★ (Transcendent) → 6★ (Transcendent ★) | Grade B → Grade S + G7 gate |
Each demerit demotes the skill by one star (floored at 0★, valid for 2★ and above only).
registry/ curated registry data and public generated catalogs
registry-for-review/ pending skill batch intake records
skill-trees/ per-user skill-tree.json files
generated-output/ ignored local scan and render output (gitignored)
docs/ docs site
src/gaia_cli/ Python CLI package
packages/cli-npm/ npm wrapper package
packages/mcp/ MCP server package
scripts/ validation, rendering, docs, and release helpers
tests/ Python test suite
.agents/skills/ Internal agent slash commands (gaia-curate, gaia-triage, etc.)
Graph is canonical. Everything else is a shadow.