| name | csb - Consensus Building |
| description | Move a group toward agreement by surfacing real disagreements, generating options, and using gradients of agreement rather than binary votes. |
| output | {"format":"prose"} |
Consensus Building
Input: $ARGUMENTS
Step 1: Surface All Positions
Before seeking agreement, map where everyone actually stands.
THE QUESTION/DECISION: [what the group needs to agree on]
POSITIONS HELD:
- Position A: [description] — Held by: [who]
- Position B: [description] — Held by: [who]
- Position C: [description] — Held by: [who]
UNSTATED POSITIONS: [any positions people may hold but haven't voiced]
Technique: Use a round-robin or anonymous written input. Do NOT start with the loudest voice or the leader's preference.
Step 2: Identify Areas of Existing Agreement
Most groups agree on more than they think. Find the common ground first.
ALREADY AGREED (things no one disputes):
1. [point of agreement]
2. [point of agreement]
3. [point of agreement]
SHARED VALUES/GOALS (even if methods differ):
1. [shared value]
2. [shared value]
PERCENTAGE OF TERRITORY ALREADY AGREED: [rough estimate]
Step 3: Identify the Real Disagreements
Surface disagreements are often proxies for deeper ones. Dig down.
SURFACE DISAGREEMENT: [what people say they disagree about]
UNDERLYING DISAGREEMENT: [what they actually disagree about]
COMMON UNDERLYING CAUSES:
- [ ] Different facts/information (fixable with data)
- [ ] Different values/priorities (requires trade-offs)
- [ ] Different risk tolerance (requires explicit discussion)
- [ ] Different time horizons (short-term vs long-term)
- [ ] Different interests/incentives (requires acknowledging)
- [ ] Lack of trust (requires process, not arguments)
REAL DISAGREEMENTS TO RESOLVE:
1. [disagreement] — Type: [facts/values/risk/time/interests/trust]
2. [disagreement] — Type: [facts/values/risk/time/interests/trust]
Step 4: Generate Options
Move from "my position vs. your position" to "what options serve our shared goals?"
OPTIONS GENERATED:
1. [option] — Addresses concerns of: [who]
2. [option] — Addresses concerns of: [who]
3. [option — hybrid/creative] — Addresses concerns of: [who]
OPTION GENERATION TECHNIQUES USED:
- [ ] "What would satisfy both sides?"
- [ ] "What if we did X on a trial basis?"
- [ ] "What if we split the decision?" (do A for context 1, B for context 2)
- [ ] "What's the smallest version we could all support?"
Step 5: Use Gradients of Agreement
Binary yes/no kills nuance. Use a gradient instead.
GRADIENT OF AGREEMENT (for each option, each person rates):
1 = Wholeheartedly support
2 = Support with minor reservations
3 = Can live with it (will not block)
4 = Have significant concerns (but won't block if heard)
5 = Cannot support (would block — must explain why)
OPTION [X] RATINGS:
- [Person]: [1-5] — Note: [any conditions]
- [Person]: [1-5] — Note: [any conditions]
THRESHOLD FOR CONSENSUS: No 5s, and average below [3.0]
If anyone rates a 5, their concerns must be addressed before proceeding. This is not a veto — it is a signal that the option needs modification.
Step 6: Lock In Agreement and Document
Agreement that isn't documented doesn't exist.
CONSENSUS REACHED:
DECISION: [the agreed-upon path]
CONDITIONS: [any conditions or modifications required to get consensus]
CONCERNS NOTED: [concerns that were heard but did not block]
REVIEW DATE: [when the group will revisit to see if concerns materialized]
COMMITMENTS:
- [Person] will: [action] by [date]
- [Person] will: [action] by [date]
WHAT "DISAGREE AND COMMIT" MEANS HERE:
[Anyone who rated 3-4 agrees to support the decision publicly and give it a fair chance through the review date]
Integration
Use with:
/tfac -> Facilitate the consensus-building session/mtgd -> Structure the meeting around the consensus process/cmp -> Compare the final options side-by-side before rating
