| name | conflict - Conflict Analysis |
| description | Map the structure of interpersonal or organizational conflict — identify the real disagreement (often hidden behind the stated one), each party's actual interests vs positions, and paths to resolution that address interests rather than positions. |
| output | {"format":"prose"} |
CONFLICT - Conflict Analysis
Input: $ARGUMENTS
Interpretations
Before executing, identify which interpretation matches the user's input:
Interpretation 1 — Active conflict resolution: The user is currently in a conflict and wants to understand its structure well enough to find a path through it — not "who's right" but "what's actually going on and what would resolve it."
Interpretation 2 — Conflict diagnosis: The user observes a conflict between others (team members, departments, stakeholders) and wants to understand its real structure so they can intervene effectively.
Interpretation 3 — Conflict prevention: The user anticipates a conflict (upcoming decision, change, announcement) and wants to map where it will emerge and how to preempt it.
If ambiguous, ask: "I can help with resolving a conflict you're in, diagnosing one you're observing, or preventing one you see coming — which fits?"
If clear from context, proceed with the matching interpretation.
Corruption Pre-Inoculation
Siding with the user is the default failure mode. The user is one party to the conflict (or sympathizes with one party). If your analysis consistently makes one side reasonable and the other unreasonable, you are advocating, not analyzing. Both parties are locally rational. Find why.
Full protocol: _shared/corruption-pre-inoculation.md
Core Principles
-
Positions are not interests. What people SAY they want (positions) is almost never the full picture. Underlying interests — security, recognition, autonomy, fairness, belonging — drive the positions. Two incompatible positions can often be resolved by addressing the underlying interests differently.
-
The stated conflict is rarely the real conflict. Arguments about budgets are often about respect. Arguments about process are often about trust. Arguments about strategy are often about identity. If you only address the surface, the conflict will resurface elsewhere.
-
Conflict has structure. It is not chaos. Every conflict has: parties, positions, interests, power dynamics, history, triggers, and constraints. Map the structure and resolution paths become visible.
-
Both sides are locally rational. If one party seems unreasonable, you haven't understood their constraints, information, or values yet. The analysis is not complete until both sides make sense from inside their own frames.
-
Power asymmetry shapes everything. Who can walk away? Who has alternatives? Who controls resources? Who sets the agenda? Power doesn't make someone wrong, but it determines which resolution paths are actually available.
-
Emotions are data, not noise. Anger signals a boundary violation. Fear signals a perceived threat. Contempt signals a status claim. Hurt signals a betrayal narrative. Read the emotions as information about what's at stake.
Phase 1: CONFLICT MAPPING
Step 1: Identify the Conflict Surface
STATED CONFLICT: [what the parties say they disagree about]
TRIGGER: [what specific event or action brought it to the surface]
PARTIES: [who is involved — name/role each as P1, P2, P3...]
DURATION: [how long has this been building?]
ESCALATION LEVEL: [disagreement / argument / fight / war / cold war]
Step 2: Map Each Party's Position and Interests
For EACH party:
PARTY: [P1 — name/role]
POSITION (what they say they want):
- [stated demand or stance]
INTERESTS (why they want it — the needs beneath the position):
- [I1] [interest — categorize as: security / recognition / autonomy / fairness / belonging / competence / resources]
- [I2] [interest]
- [I3] [interest]
CONSTRAINTS (what limits their flexibility):
- [things they can't or won't give up, and why]
INFORMATION (what they know/believe):
- What they know that the other party doesn't: [asymmetric info]
- What they believe that may be wrong: [mistaken beliefs]
- What they assume about the other party's motives: [attribution]
EMOTIONAL STATE:
- Primary emotion: [anger / fear / hurt / contempt / frustration]
- What it signals: [boundary violation / threat / betrayal / status claim]
POWER:
- Sources of power: [authority / expertise / alternatives / relationships / information]
- BATNA (Best Alternative To Negotiated Agreement): [what they do if no resolution]
Step 3: Find the Real Conflict
The stated conflict is the surface. Dig for what's underneath.
LAYER ANALYSIS
==============
SURFACE LAYER: [what they're arguing about — the stated issue]
PROCESS LAYER: [how they're arguing — communication patterns, who gets heard]
RELATIONSHIP LAYER: [what the conflict means about their relationship — trust, respect, status]
IDENTITY LAYER: [what the conflict threatens about who they are]
THE REAL CONFLICT: [1-2 sentences — what is this ACTUALLY about?]
EVIDENCE: [why you believe this is the real conflict — behavioral clues, emotional intensity, pattern of escalation]
Phase 2: DYNAMICS ANALYSIS
Step 4: Map the Conflict Dynamics
ESCALATION PATTERN:
- [A] [P1 does/says X] → [P2 interprets as Y] → [P2 responds with Z]
- [B] [P2 does/says X] → [P1 interprets as Y] → [P1 responds with Z]
- [C] [The cycle: how A and B feed each other]
STRUCTURAL DRIVERS (things keeping the conflict alive):
- [S1] [driver — e.g., shared resource, reporting structure, unclear ownership]
- [S2] [driver]
HISTORICAL CONTEXT:
- [Past events that shaped current interpretations]
- [Unresolved prior conflicts that feed this one]
THIRD-PARTY EFFECTS:
- [Who else is affected? Who is pulled in? Who benefits from the conflict continuing?]
Step 5: Interest Overlap Analysis
INTEREST COMPARISON
===================
SHARED INTERESTS (both parties want this):
- [interest that both parties share, even if they don't realize it]
- [interest]
COMPATIBLE INTERESTS (different but not conflicting):
- P1 wants [X], P2 wants [Y] — these don't conflict because [reason]
GENUINELY CONFLICTING INTERESTS (zero-sum):
- P1 wants [X], P2 wants [NOT-X] — can't both have it because [reason]
- Is this ACTUALLY zero-sum, or does it just look that way? [analysis]
KEY FINDING: [What percentage of the conflict is genuinely zero-sum vs. resolvable through interest-based approaches?]
Phase 3: RESOLUTION PATHS
Step 6: Generate Resolution Options
For each genuinely conflicting interest, generate options:
RESOLUTION OPTIONS
==================
[R1] [option name]
Addresses P1's interest in: [which interests]
Addresses P2's interest in: [which interests]
Requires: [what has to happen]
Risk: [what could go wrong]
Feasibility: [HIGH / MEDIUM / LOW — given power dynamics and constraints]
[R2] [option name]
...
[R3] [option name — the creative option that reframes the conflict]
...
Rules for generating options:
- At least one option must address the REAL conflict, not just the stated one
- At least one option must be creative (changes the frame, not just the terms)
- No option should require one party to simply capitulate
- Each option must reference specific interests from the mapping
Step 7: Compile the Analysis
CONFLICT ANALYSIS: [situation]
==============================
SURFACE vs. REAL:
Stated: [what they say it's about]
Actual: [what it's really about]
PARTY SUMMARY:
P1: Wants [position] because [interests]. Fears [threat]. Power: [assessment].
P2: Wants [position] because [interests]. Fears [threat]. Power: [assessment].
ZERO-SUM PERCENTAGE: [how much is genuinely irreconcilable]
ESCALATION RISK: [HIGH / MEDIUM / LOW — what happens if nothing changes]
RECOMMENDED PATH: [R-number] because [derived from interest analysis]
FIRST MOVE: [specific, concrete action — not "have a conversation" but what to say/do]
WHAT MUST BE ADDRESSED FOR ANY RESOLUTION:
- [non-negotiable element — from the real conflict layer]
WHAT WILL MAKE IT WORSE:
- [specific actions to avoid — derived from escalation pattern]
CONFIDENCE: [HIGH / MEDIUM / LOW]
WEAKEST ASSUMPTION: [what part of this analysis is most likely wrong]
Failure Modes
| Failure | Signal | Fix |
|---|
| Taking sides | One party is reasonable, the other irrational | Find the rationality in the "irrational" party. Both sides make sense from inside. |
| Surface-only analysis | Resolution addresses stated positions, not underlying interests | Go deeper. Ask: if they got exactly what they say they want, would they actually be satisfied? |
| False equivalence | "Both sides have a point" without specificity | Name WHICH points each side has. Vague balance is not analysis. |
| Power blindness | Resolution assumes equal negotiating position | Map power explicitly. Who can walk away? Who controls resources? |
| Resolution by fiat | "They should just communicate better" | Communication is a tactic, not a strategy. What specifically should be communicated, and why would it work? |
| Ignoring structural drivers | Treating conflict as purely interpersonal when systems cause it | Check: would different people in these roles have the same conflict? If yes, it's structural. |
Depth Scaling
| Depth | Parties Analyzed | Interest Layers | Resolution Options | Min Findings |
|---|
| 1x | 2 | 2 (surface + real) | 2 | 6 |
| 2x | 2-3 | 3 (surface + process + real) | 3 | 10 |
| 4x | 3-4 | 4 (all layers) | 5 | 18 |
| 8x | 4+ | 4 (all layers, each party) | 7 | 30 |
Default: 2x. These are floors.
Pre-Completion Checklist
Integration
- Use from:
/empth (understand each party first), /emotion (when feelings are driving the conflict)
- Routes to:
/persua (how to communicate the resolution), /trust (if trust repair is needed), /empth (if one party's perspective is poorly understood)
- Differs from:
/cmp (comparing options vs. mapping conflict structure), /dcp (making YOUR decision vs. resolving a dispute between parties)
- Complementary:
/empth (model each party), /trust (assess trust damage and repair), /aw (stress-test your conflict model)
