{"full_name":"0xNyk/council-of-high-intelligence","owner":"0xNyk","name":"council-of-high-intelligence","url":"https://github.com/0xNyk/council-of-high-intelligence","description":"Structured multi-perspective deliberation for hard decisions. Run full councils, focused triads, or duo debates across Claude Code, Codex, Gemini CLI, and OpenCode.","homepage":"https://www.nyk.dev/oss/council-of-high-intelligence","topics":[],"language":"Shell","license":"MIT","avatar":"https://avatars.githubusercontent.com/u/93952610?u=ca60fa25d8f74d8103b68dfbb2fe3eeebfd8402d&v=4","created_at":"2026-03-02T08:47:09Z","pushed_at":"2026-11-09T05:21:24Z","metrics":{"stars":4831,"forks":455,"watchers":20,"issues":24,"prs":64,"releases":3,"issues_open":null,"issues_closed":null,"prs_open":null,"prs_closed":null,"prs_merged":null,"discussions":0},"growth":{"day":{"stars":10,"forks":-1,"base_day":"2026-11-09"},"week":{"stars":92,"forks":8,"base_day":"2026-11-03"},"month":{"stars":291,"forks":25,"base_day":"2026-10-11"}},"history":{"dates":["2026-09-30","2026-10-01","2026-10-02","2026-10-03","2026-10-04","2026-10-05","2026-10-06","2026-10-07","2026-10-08","2026-10-09","2026-10-10","2026-10-11","2026-10-12","2026-10-13","2026-10-14","2026-10-15","2026-10-16","2026-10-17","2026-10-18","2026-10-19","2026-10-20","2026-10-21","2026-10-22","2026-10-23","2026-10-24","2026-10-25","2026-10-26","2026-10-27","2026-10-28","2026-10-29","2026-10-30","2026-10-31","2026-11-01","2026-11-02","2026-11-03","2026-11-04","2026-11-09","2026-11-10"],"stars":[4449,4458,4487,4498,4507,4512,4527,4528,4536,4532,4536,4540,4543,4556,4570,4579,4593,4617,4625,4643,4652,4658,4665,4677,4683,4686,4691,4696,4700,4710,4715,4722,4727,4730,4739,4761,4821,4831],"forks":[421,422,423,425,426,428,428,428,428,428,430,430,429,429,429,431,431,433,435,437,439,443,447,456,456,443,444,444,444,444,445,444,444,445,447,450,456,455],"watchers":[21,21,21,21,21,21,21,21,21,20,20,20,20,20,20,20,20,21,21,21,21,21,21,21,21,21,21,21,21,21,21,21,20,20,20,20,20,20],"issues_open":[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null],"prs_open":[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null]},"boards":{},"same_owner":[{"full_name":"0xNyk/awesome-hermes-agent","name":"awesome-hermes-agent","boards":{}}],"issues":[{"number":88,"title":"bug:  --no-auto-route is not default","comments":2,"created_at":"2026-11-06T19:00:26Z"},{"number":34,"title":"proposal: add support for modular python agent frameworks (Agno, LangGraph, etc.)","comments":2,"created_at":"2026-05-26T19:41:21Z"},{"number":71,"title":"feat: add Abraham Wald as 19th council persona (survivorship bias and selection-corrected inference)","comments":1,"created_at":"2026-08-01T22:11:08Z"},{"number":28,"title":"tracking: 2026 MAD literature enhancement candidates (E1–E8)","comments":1,"created_at":"2026-05-18T13:11:00Z"},{"number":84,"title":"codex_exec seats fail outside a git repo and can hang waiting on stdin","comments":0,"created_at":"2026-09-24T22:12:55Z"},{"number":83,"title":"openai_compatible_api seats: max_tokens 1200 makes reasoning models return empty answers","comments":0,"created_at":"2026-09-24T22:12:54Z"},{"number":82,"title":"openai_compatible_api seats: max_tokens 1200 makes reasoning models return empty answers","comments":0,"created_at":"2026-09-24T21:43:37Z"},{"number":81,"title":"proposal: evaluate draft/critique/revise as a first step toward adaptive workflow routing","comments":0,"created_at":"2026-09-05T10:26:34Z"}],"discussions":[],"latest_release":{"name":"v1.2.0","tag":"v1.2.0","published_at":"2026-07-04T07:13:48Z","body":"\n### Added\n- **Claude Code plugin-marketplace support** (#45) — `.claude-plugin/plugin.json` + `marketplace.json`; install with `/plugin marketplace add 0xNyk/council-of-high-intelligence` then `/plugin install council@council-of-high-intelligence`. Root `SKILL.md` remains the single source of truth via a `skills/council/` symlink; a new Asset Resolution section resolves agents/scripts/configs from either the install.sh layout or `${CLAUDE_PLUGIN_ROOT}`.\n- **Confidence-weighted vote tally** (#44) — a member's vote weight is now base weight (1.0, or 1.5× domain seat) × confidence factor from their `STANCE:` line (`high 1.0 / med 0.75 / low 0.5`). `W_total` stays at full base weights, so a hesitant council raises its own consensus bar and escalates to the user rather than forcing a verdict (Roundtable Policy, arXiv:2509.16839; ConfMAD, arXiv:2509.14034).\n- **Per-member `reasoning_method` frontmatter** (E3 / DMAD, arXiv:2410.12853) (#44) — all 18 personas declare a distinct reasoning method (elenchic-questioning, tail-stress-testing, causal-loop-mapping, …) and Round 1 prompts direct each member to reason via their method: method diversity, not just persona diversity.\n- **Project-level `./.council.yaml` overrides** (E7) (#44) — pin `profile`/`triad`/`members`/`chairman`/`models`/`no_auto_route` per project; CLI flags always win.\n- Standard repo furniture (#42) — `.gitignore`, `SECURITY.md`, `FUNDING.yml`, bug/feature/provider issue templates, and a PR template with a three-file protocol-parity checklist.\n- **Structured stance voting + deterministic weighted tie-break.** The final round (full Round 3 / quick Round 2) now requires each member to emit a machine-parseable `STANCE: <option> | CONFIDENCE: … | DEALBREAKER: …` line, so consensus is a counted weighted tally rather than a prose impression. STEP 6 specifies the exact math: `W_option ≥ (2/3) × W_total`, with the on-domain seat carrying 1.5×. New **Vote Tally** field in the full and quick verdict templates records `option → weight` and which seat carried the weight. Mirrored in `SKILL.codex.md`.\n- Cursor CLI provider support (`cursor_cli` archetype) — sixth dispatch path alongside subagent / codex_exec / gemini_cli / ollama_run / openai_compatible_api. Auto-detected via the `cursor-agent` binary; members run headless and read-only (`cursor-agent -p --mode ask --model <id>`). Cursor is a model aggregator (GPT-5.x / Claude / Gemini / Grok through one CLI), so routing treats it as a single provider for spread and steers diversity seats to cross-family models to avoid duplicating Anthropic bias. New `configs/provider-model-slots.cursor.example.yaml`, `cursor_cli` tiers in `configs/auto-route-defaults.yaml`, and `cursor_cli` as a valid `--chairman` tag.\n\n### Changed\n- **Domain-weight seat (1.5×) is now designated at panel selection (STEP 0), before any positions exist** — previously chosen at tie-break time, which let the coordinator pick the heavyweight after seeing votes. Locking it up front removes that nudge.\n- README \"Enforcement Mechanisms\" rewritten to state the actual forcing function (bounded round budget + anti-recursion guards) and that genuine splits escalate to the user instead of being forced into consensus.\n- **CI drift guards are now real** (#42) — `council-simulation-checklist.sh` executes in CI (it never ran before), extended with full `SKILL.gemini.md` parity checks and a three-file STANCE/Vote-Tally/1.5×/2-3-threshold parity gate; markdownlint's variadic `--disable` was missing its trailing `--` and silently linted zero files — rule selection now lives in `.markdownlint.json` and covers all doc files.\n- `install.sh` derives the generated `gemini-extension.json` version from CHANGELOG instead of a hardcoded string; model IDs reconciled across configs/docs (`gemini-3-pro`, `gpt-5.4` generation); README/CLAUDE.md document the `--gemini` install path and marketplace install.\n\n### Fixed\n- **`SKILL.gemini.md` drift regression** (#42) — the Gemini coordinator was miss"},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-10T09:35:25Z","history_start":"2026-09-06"}