{"_id":"@demi8-patch/model-council","name":"@demi8-patch/model-council","dist-tags":{"latest":"0.1.0"},"versions":{"0.1.0":{"name":"@demi8-patch/model-council","version":"0.1.0","description":"Multi-perspective AI deliberation skill — convene a council of analytical perspectives, cross-validate findings, and surface consensus, unique discoveries, and conflicts. Inspired by Perplexity's Model Council.","keywords":["pi-package","pi-skill","model-council","multi-perspective","deliberation"],"license":"MIT","author":{"name":"demi8-patch"},"repository":{"type":"git","url":"git+https://github.com/demi8-patch/model-council.git"},"scripts":{"lint":"grep -rn 'TBD\\|TODO\\|FIXME' skills/model-council/ 2>/dev/null && (echo 'FAIL: unresolved markers found' && exit 1) || echo 'PASS: clean'","verify:structure":"grep -q 'Phase 1.*Convene' skills/model-council/SKILL.md && grep -q 'Phase 2.*Collect' skills/model-council/SKILL.md && grep -q 'Validate-and-Recover' skills/model-council/SKILL.md && grep -q 'Phase 3.*Synthesize' skills/model-council/SKILL.md && grep -q 'Phase 4.*Present' skills/model-council/SKILL.md && echo 'PASS: 5 phases' || (echo 'FAIL: SKILL.md missing phases' && exit 1)","verify:perspectives":"for f in skills/model-council/perspectives/skeptic.md skills/model-council/perspectives/systems-thinker.md skills/model-council/perspectives/evidence-auditor.md skills/model-council/perspectives/optimist.md skills/model-council/perspectives/devil-advocate.md; do grep -q 'confidence rubric' \"$f\" || { echo \"FAIL: $f missing confidence rubric\"; exit 1; }; done && echo 'PASS: all 5 perspectives have bodies'","verify:synthesizer":"grep -q 'You are the Council Synthesizer' skills/model-council/references/synthesizer-prompt.md && grep -q 'Citation contract' skills/model-council/references/synthesizer-prompt.md && grep -q 'Council status:' skills/model-council/references/synthesis-format.md && echo 'PASS: synthesizer contract' || (echo 'FAIL: synthesizer references incomplete' && exit 1)","verify:call-counts":"grep -q '5 subagent' README.md && grep -q '7 subagent' README.md && grep -q '8 subagent' README.md && grep -q '5 subagent' skills/model-council/SKILL.md && grep -q '7 subagent' skills/model-council/SKILL.md && grep -q '8 subagent' skills/model-council/SKILL.md && echo 'PASS: call counts consistent' || (echo 'FAIL: call count mismatch' && exit 1)","verify:all":"npm run lint && npm run verify:structure && npm run verify:perspectives && npm run verify:synthesizer && npm run verify:call-counts","prepare":"husky"},"pi":{"skills":["./skills"]},"peerDependencies":{"pi-subagents":"*"},"devDependencies":{"husky":"^9.0.0","lint-staged":"^15.0.0","prettier":"^3.0.0"},"_id":"@demi8-patch/model-council@0.1.0","gitHead":"7f84c5808923ab79036c6e579a25fe3520119a68","bugs":{"url":"https://github.com/demi8-patch/model-council/issues"},"homepage":"https://github.com/demi8-patch/model-council#readme","_nodeVersion":"22.23.1","_npmVersion":"10.9.8","dist":{"integrity":"sha512-AyJ3Cuppekui7ipQVey6DnySXEmyp9rlJcyVbLCEtM15FTZ4laWHThLGsK3u8I0Aa6s3RwNFOtJNqJEnsI71yA==","shasum":"9d07c5392c9f5289090cadbb0586fab031862dae","tarball":"https://registry.npmjs.org/@demi8-patch/model-council/-/model-council-0.1.0.tgz","fileCount":91,"unpackedSize":482347,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEYCIQCxPv3FGfAzuvVcrDhX/dztKyP9co6mjVz5QosH1GmSvQIhAOrDXO3xoi0qqY63DHhr9Vujv64X5RCQcf2FhFWehlMG"}]},"_npmUser":{"name":"demi8_patch","email":"dtripath4@gmail.com"},"directories":{},"maintainers":[{"name":"demi8_patch","email":"dtripath4@gmail.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/model-council_0.1.0_1783604804089_0.878194666233042"},"_hasShrinkwrap":false}},"time":{"created":"2026-07-09T13:46:43.870Z","0.1.0":"2026-07-09T13:46:44.244Z","modified":"2026-07-09T13:46:44.463Z"},"maintainers":[{"name":"demi8_patch","email":"dtripath4@gmail.com"}],"description":"Multi-perspective AI deliberation skill — convene a council of analytical perspectives, cross-validate findings, and surface consensus, unique discoveries, and conflicts. Inspired by Perplexity's Model Council.","homepage":"https://github.com/demi8-patch/model-council#readme","keywords":["pi-package","pi-skill","model-council","multi-perspective","deliberation"],"repository":{"type":"git","url":"git+https://github.com/demi8-patch/model-council.git"},"author":{"name":"demi8-patch"},"bugs":{"url":"https://github.com/demi8-patch/model-council/issues"},"license":"MIT","readme":"# model-council\n\nMulti-perspective AI deliberation for the pi coding agent. Inspired by Perplexity's Model Council.\n\nConvene a council of analytical perspectives, cross-validate findings, and surface consensus, unique discoveries, and conflicts — so you know when to move faster with confidence, and when to dig deeper.\n\n## Install\n\n```bash\npi install npm:@demi8-patch/model-council\n```\n\nRequires `pi-subagents` (auto-installed as peer dependency on first use).\n\n## Usage\n\nInvoke the skill when facing high-stakes questions where single-perspective blind spots are costly:\n\n```\n/skill:model-council\n```\n\nOr let the agent fire it automatically — trigger words: \"council\", \"deliberate\", \"multi-perspective\", \"cross-validate\", \"second opinion\".\n\n## What you get\n\nA consolidated answer in four sections:\n\n- **Consensus** — what the council independently agreed on (with citations)\n- **Unique discoveries** — insights only one perspective caught\n- **Conflicts** — where perspectives disagree, and why\n- **Confidence** — whether to move faster or dig deeper\n\n## Cost transparency\n\nCost: approximately $0.50–$3.00 per council depending on perspective count and model choices. Full breakdown:\n\n| Path                     | Calls                                                   |\n| ------------------------ | ------------------------------------------------------- |\n| 3-perspective happy path | **5 subagent calls** (Convene + 3 Collect + Synthesize) |\n| 5-perspective deep mode  | **7 subagent calls** (Convene + 5 Collect + Synthesize) |\n| Worst-case with retries  | up to **8 subagent calls**                              |\n\nEach `subagent()` call expects 2k-8k output tokens. Suggested timeout: 120 seconds per call. The full 4-section synthesis output is capped at ~1200 words.\n\n## License\n\nMIT — see [LICENSE](./LICENSE).\n","readmeFilename":"README.md","_rev":"1-c5ffafcc2d942fd8fe32049e4fcf7a3d"}