| name | build-intelligence-map |
| model | opus |
| description | Produces a cross-tool disposition matrix, behavioral cluster analysis, and a set of named YAML disposition presets by synthesizing analyses from 3+ ingested AI system prompts. Use when: "build an intelligence map", "compare all agent behaviors", "generate disposition presets from external tools", "what can we learn from other tools", "synthesize agent analyses". |
| license | Complete terms in LICENSE.txt |
| category | wisdom-garden |
| triggers | ["build an intelligence map","compare all agent behaviors","generate disposition presets from external tools","synthesize agent analyses"] |
| tier | 1 |
| agents | ["primary"] |
| tool_dependencies | ["file_system"] |
| inputs | [{"name":"agent_analyses","type":"string[]","description":"Paths to 3+ ingested AI system prompt analysis files to synthesize","required":true}] |
| outputs | [{"name":"intelligence_map","type":"ref","format":"cas-ref","description":"Cross-tool disposition matrix, behavioral cluster analysis, and named YAML disposition presets"}] |
Build Intelligence Map
Meta-skill that synthesizes multiple agent behavioral analyses into strategic intelligence.
I. Philosophy
Individual agent analysis reveals how one tool thinks. The intelligence map reveals the design space — the full range of behavioral choices the industry has explored, and where Dojo sits within that landscape. This is not competitive intelligence in the traditional sense; it is cartography of agent behavioral design.
The map serves three purposes: (1) validate that Dojo's defaults are well-positioned, (2) identify behavioral combinations no tool has tried that might be valuable, and (3) build a library of disposition presets that agents can switch between based on task context.
II. When to Use
- After running
ingest-system-prompt + analyze-agent-behavior on 3+ tools
- When evolving the ADA disposition system with new presets
- When a user asks "what are all the tools doing differently?"
- During strategic planning for agent behavior features
Do NOT use with fewer than 3 analyzed tools (insufficient data for meaningful comparison) or when the goal is to analyze a single tool (use analyze-agent-behavior directly).
III. Workflow
Step 1: Gather Analyses
Query MemoryStore for all entries with EntryType: "system_prompt":
- Retrieve each stored analysis (disposition YAML + evidence catalog)
- Validate: at least 3 tools with completed analyses
- Sort by analysis date (most recent first)
Step 2: Build the Disposition Matrix
Construct a tool x field matrix:
| pacing | depth | tone | initiative | validation | error_handling |
------------|------------|------------|--------------|------------|-------------|----------------|
Cursor | responsive | functional | professional | responsive | thorough | retry |
Windsurf | rapid | surface | conversational| proactive | spot-check | log-and-cont. |
Copilot | measured | thorough | professional | responsive | thorough | escalate |
Claude Code | measured | thorough | professional | responsive | thorough | log-and-cont. |
Dojo Default| measured | thorough | professional | responsive | thorough | log-and-cont. |
For each cell, include the confidence score. Color-code or annotate cells where confidence < 0.6.
Step 3: Identify Clusters
Group tools by behavioral similarity:
- Conservative cluster: High validation, escalate errors, deliberate pacing
- Velocity cluster: Rapid pacing, surface depth, autonomous initiative
