Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
Pre-indexed code knowledge graph, auto syncs on code changes, for Claude Code, Codex, Gemini, Cursor, OpenCode, AntiGravity, Kiro, and Hermes Agent — fewer tokens, fewer tool calls, 100% local
npx @colbymchenry/codegraphA hand-picked collection of the finest of resources for the most awesome of agents, Claude Code, the undisputed champion of coding companions, from the unstoppable team at Anthropic PBC. A delectable showcase of top tier skills, ambidextrous agents, scintillating status lines, top notch developer tooling, and also we have plugins
These intent pages connect this repository to workflow-first and comparison-first discovery routes.
Carries strong trust indicators from repository metadata
2033 GitHub stars recorded
🆕 v9.56.1 — Harden release guardrails and close review findings.
Default roster: Claude Opus 5 leads architecture, planning, security reasoning, and final judgment; GPT-5.6 Sol is the independent implementation/review peer; Claude Sonnet 5 is the standard Claude seat; Fable 5 remains an opt-in judgment escalation. Existing model pins and provider configuration still win. See the routing strategy.
/octo:model-config # inspect or override the frontier roster OCTOPUS_OPUS5_AUTO_XHIGH=1 # opt in to automatic xhigh Opus 5 phases OCTOPUS_OPUS_MODEL=claude-fable-5 # explicitly opt in to Fable 5
🆕 v9.41 — Multi-LLM Council.
/octo:councilruns a structured 3/5/7-persona deliberation across Claude, Codex, Gemini, and OpenCode with goal modes (advice,decision,plan,implement,review), styles (balanced,adversarial,red-team,executive,implementation), benchmark-aware role routing, quorum + critical-veto gates, budget caps, and gated worktree handoff for approved plans. Use it when one model's opinion isn't enough./octo:council --goal decision --style adversarial "Should this service stay monolithic?" /octo:council --goal implement --implement plan-only "Refactor the auth flow"
| Version | Best Features |
|---|---|
| v9.56.1 (new) | Harden release guardrails and close review findings. |
| v9.50 | Claude Code 2026 compatibility layer — routines manifest (schedule + GitHub-event automations), SubagentStop quality/cost gate, /octo:usage cost attribution, worktree.bgIsolation opt-out, Claude Agent SDK seat (introduced with Opus 4.8 and now following the current Opus 5 default), starter skills pack, /plugin browse manifest with projected context cost. |
| v9.41 | /octo:council promoted to first-class workflow — structured multi-LLM deliberation with goal modes, adversarial/red-team styles, benchmark-aware persona routing, quorum and critical-veto gates, budget preflight, and gated worktree handoff for approved implementation plans. |
| v9 (current) | Up to 10 external provider integrations (Codex, Gemini, Antigravity CLI, Copilot, Qwen, Ollama, Perplexity, OpenRouter, OpenCode, and Grok) alongside the Claude Code host. Structured provider debates and configurable multi-LLM councils. Smart router — just say what you need. Agent summary tables show which providers actually contributed. Provider-aware prompt preflight prevents silent oversize failures. Research breadth modes fan out light, standard, or exhaustive investigations. Setup aliases and fuzzy /octo:* corrections reduce command friction. Discipline mode with 8 auto-invoke gates. Two-stage review. Circuit breakers with automatic provider recovery. Cursor + OpenCode + Codex cross-compatibility. Token compression: bin/octo-compress pipe + auto PostToolUse hook save ~7,300 tokens/session. PostCompact context recovery. bin/octopus CLI. 182 Claude Code capability flags through v2.1.219, including Opus 5, Sonnet 5, and dynamic workflow awareness. |
| v8 | Multi-LLM code review with inline PR comments. Parallel workstreams in isolated git worktrees. Reaction engine — auto-responds to CI failures. 32 specialized personas. Dark Factory autonomous pipeline. |
providers.json pins remain unchanged; OCTOPUS_LEGACY_ROLES=1 restores the pre-frontier role mapping.CLAUDE_SDK_API_KEY, OCTOPUS_CLAUDE_SDK_MODEL, OCTOPUS_CLAUDE_SDK_MAX_TOKENS, OCTOPUS_CLAUDE_SDK_ALLOWED_MODELS, OCTOPUS_CLAUDE_SDK_CONTEXT_BUDGET.OCTOPUS_FABLE5_MODE (auto/off/on), OCTOPUS_FABLE5_NO_RETRY. Guards auto-enable only when you pin claude-fable-5.OCTOPUS_LEGACY_ROLES=1.| Method | Codex | Gemini | Antigravity | Claude |
|---|---|---|---|---|
| OAuth/subscription (recommended) | codex login — included in ChatGPT subscription | Google account — included in AI subscription | agy auth — included with Antigravity access | Built into Claude Code |
| API key | OPENAI_API_KEY — per-token billing | GEMINI_API_KEY — per-token billing | n/a | Built into Claude Code |
OAuth users pay nothing beyond their existing subscriptions. Qwen is the exception: its free OAuth tier ended on 2026-04-15, so use QWEN_API_KEY or Coding-Plan (OPENAI_API_KEY + OPENAI_BASE_URL).
Illustrative token-only estimates, using standard global API rates checked 2026-07-27: GPT-5.6 Sol $5/$30, Gemini 3.1 Pro Preview $2/$12 for prompts through 200K tokens and $4/$18 for prompts over 200K, Sonar Pro $3/$15, and Opus 5 $5/$25 per million input/output tokens. The ranges assume roughly 90% input and 10% output tokens, standard (not batch, flex, priority, or fast) processing, no cache discounts, and a representative mix of those models. OAuth/subscription seats (Codex via ChatGPT, Gemini via Google account, Antigravity, Copilot) bill nothing extra; Ollama is free.
| v7 | Double Diamond workflow. Multi-provider dispatch. Quality gates and consensus scoring. Configurable sandbox modes. |
The table excludes provider tool charges. Sonar Pro adds a request fee of $6-$14 per 1,000 requests depending on search-context size; Gemini search grounding can add query fees after its included allowance. Long-context and provider-specific rate rules can push large runs above these bounds, so check the linked rate cards before material spend.
| Run | Typical volume | Illustrative API token cost (tool fees excluded) |
|---|---|---|
| Single probe / quick question (one provider) | 5-20K tokens | $0.01-0.20 |
| Debate (2-3 providers, multi-round) | 30-80K tokens | $0.20-1.00 |
| Council (4-6 seats + synthesis) | 60-150K tokens | $0.50-2.50 |
| Full embrace (4 phases, multi-provider) | 150-400K tokens | $1.00-6.00+ |
Before an expensive run, /octo:costs shows a session cost projection; after runs, /octo:usage breaks down actual spend per provider and skill. Anything projected over $1 is called out before dispatch.
Everything except multi-AI features. You get all 32 personas, structured workflows, smart routing, context detection, and every skill. Multi-AI orchestration (parallel analysis, debate, consensus) activates when external providers are configured.
CLAUDE_SDK_API_KEY | unset | Enables the claude-sdk provider seat (Claude Agent SDK; Opus 5, 1M context) |
OCTOPUS_CLAUDE_SDK_MODEL | claude-opus-5 | Model for the claude-sdk seat |
OCTOPUS_CLAUDE_SDK_MAX_TOKENS | 32768 | Max output tokens for the claude-sdk seat (SDK CLI path) |
OCTOPUS_CLAUDE_SDK_ALLOWED_MODELS | unset | Comma-separated model allowlist for the claude-sdk seat |
OCTOPUS_CLAUDE_SDK_CONTEXT_BUDGET | 1000000 | Prompt context budget for claude-sdk agent types |
OCTOPUS_WORKTREE_BG_ISOLATION | true | Set false to skip background worktree cloning (fast direct-edit runs) |
OCTOPUS_SUBAGENT_GATE_STRICT | false | Set true to let the SubagentStop gate block malformed council verdicts and low-quality summaries |
OCTOPUS_SUBAGENT_MIN_QUALITY | 0 | Quality floor (0-100) enforced by the gate in strict mode; 0 disables |