From sentinel
Collective decision protocol. Generates individualized pre-meeting scoring sheets, structures the meeting agenda by divergence, and facilitates group discussion to surface unique information and prevent groupthink. This is NOT /sentinel run by multiple people — it is a fundamentally different protocol that treats group dynamics as the primary risk variable.
npx claudepluginhub jamon8888/cc-suite --plugin Sentinelsonnet<example> user: "/sentinel-group 4 partners evaluating whether to open a Berlin office" assistant: "[Runs triage for group context -> generates Phase 1 scoring sheets for 4 participants -> structures Phase 2 meeting agenda -> activates noise-calculator to brief on expected variance -> outputs complete facilitation protocol]" </example> # /sentinel-group — Collective Decision Protocol ## Core premise Group decisions have a different failure mode than individual decisions. The enemy is not individual bias — it is **process contamination**: information that exists in the room but never surf...
Share bugs, ideas, or general feedback.
Group decisions have a different failure mode than individual decisions. The enemy is not individual bias — it is process contamination: information that exists in the room but never surfaces, preferences that exist but are suppressed, disagreements that exist but are smoothed over.
The protocol does not try to make individuals less biased. It structures the process so that group dynamics produce signal rather than noise.
Apply the inline triage formula (same as /sentinel Step 1), with group = true forced.
This adds +8.0 to the score and activates activate_group_facilitator.
Additional group parameters to assess from context:
Use structure-builder logic to identify 3-5 dimensions. For group decisions, add a mandatory dimension: "Process legitimacy" (will the people who didn't get their way accept the outcome as fair?) This dimension is always relevant in group decisions and almost always omitted.
Activate group-facilitator agent to generate individualized sheets.
Each sheet contains:
Format the sheets for distribution — plain text, copy-paste ready. Label them Sheet A, Sheet B, etc. (not by name, to reduce anchoring to person).
Before the meeting, provide the facilitator with:
Structure:
After the meeting, record:
Read data/decision-ledger.json
Append group decision entry with metadata: n_participants, seniority_spread, individual positions, group outcome, confidence levels, review_date.
Write data/decision-ledger.json
/sentinel-group output:
📋 PHASE 1 — Individual Scoring Sheets (distribute before meeting)
[Sheet A — copy-paste ready]
[Sheet B — copy-paste ready]
[...]
📊 NOISE BRIEFING — What to expect from score variance
[threshold interpretation]
🗓️ PHASE 2 — Meeting Facilitation Guide
[step-by-step agenda]
⚠️ FACILITATION WARNINGS
[seniority spread risks]
[prior alignment risks]
[courtesy bias risks]
Always respond in the user's language.