Produces a behavioral architecture map grouping components by action-verb clusters (CONVERSE, REASON, REMEMBER, etc.) that reveals capabilities, gaps, and cross-cutting concerns invisible to directory trees. Use when: 'what does this app actually do', 'walk me through the features', 'plan a refactor', 'audit feature coverage', 'explain this system to someone new'.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
Produces a behavioral architecture map grouping components by action-verb clusters (CONVERSE, REASON, REMEMBER, etc.) that reveals capabilities, gaps, and cross-cutting concerns invisible to directory trees. Use when: 'what does this app actually do', 'walk me through the features', 'plan a refactor', 'audit feature coverage', 'explain this system to someone new'.
category
system-health
triggers
["what does this app actually do","walk me through the features","plan a refactor","audit feature coverage","explain this system to someone new"]
tier
1
agents
["primary"]
tool_dependencies
["file_system"]
inputs
[{"name":"repo_path","type":"string","description":"Path to the repository or codebase to analyze","required":true}]
outputs
[{"name":"behavioral_map","type":"string","description":"Behavioral architecture map grouping components by action-verb clusters revealing capabilities, gaps, and cross-cutting concerns"}]
Semantic Clusters Skill
Version: 1.0
Created: 2026-02-08
Author: Cruz + Manus (Cowork)
Origin: Elevated from the repo-status skill's semantic clusters reference, which was codified from a live strategic-scout audit of Dojo Genesis (13 clusters, 333 Go files, 416 frontend files mapped).
Lineage: strategic-scout (exploration) → repo-status (formalization) → this standalone skill.
I. The Philosophy: Behavior Over Location
Every codebase has two architectures:
The filesystem architecture — where files live. Directories, packages, modules. This is what ls shows you.
The behavioral architecture — what the system does. Capabilities that cross-cut directories, features that span layers.
Most people only see architecture #1. They think in terms of frontend/, backend/, utils/. But understanding comes from architecture #2 — the verbs.
A "chat" feature isn't in one directory. It's a component in the frontend, a handler in the backend, a state engine in a context, a streaming service, an SSE connection, and a set of tools. These parts live in 6 different directories. But they all serve one verb: CONVERSE.
Semantic clusters make the behavioral architecture visible. Each cluster is named with an action verb, and every significant component in the codebase maps to one (sometimes two) clusters. The result is a map of what the system can do — not just where its files happen to be.
II. When to Use This Skill
Exploring a new codebase: Before diving into files, map the behavioral capabilities to build your mental model.
Explaining a system: Clusters make better explanations than directory trees because they answer "what does it do?" not "where are the files?"
Planning a refactor: Clusters reveal which components serve the same capability. Refactoring within a cluster is safer than across clusters.
Auditing feature coverage: Clusters expose gaps — capabilities the system lacks or has only partially implemented.
Identifying architectural confusion: If a component maps to 3+ clusters, it's probably doing too much. If a directory has components in 5 different clusters, it may need restructuring.
Writing status documents: Section 4 of the status-template skill uses clusters. This skill provides the methodology.
III. The Starter Verbs
These 13 action verbs cover most software systems. Read references/verb-catalog.md for detailed definitions, component examples, and signals for each.
Verb
What It Means
Example Systems That Need It
CONVERSE
Real-time communication with users
Chat apps, messaging, support tools
REASON
Thinking, planning, deciding
AI agents, rule engines, recommendation systems
REMEMBER
Storing and recalling knowledge
Knowledge bases, caching, memory systems
OBSERVE
Watching and reporting
Monitoring, analytics, logging, tracing
LEARN
Adapting based on feedback
Calibration, A/B testing, preference learning
ACT
Executing side effects
Tool systems, API calls, file operations, cron
PROTECT
Enforcing boundaries
Auth, encryption, rate limiting, validation
CONNECT
Integrating externally
Plugins, APIs, webhooks, bots, OAuth
PRESENT
Rendering UI
Shells, layouts, component libraries, themes
PERSIST
Storing data durably
Databases, migrations, ORMs, caches
BUILD
Building, testing, shipping
CI/CD, Docker, scripts, test suites
THINK
Meta-cognition about itself
Skills, prompts, documentation, retrospectives
ORCHESTRATE
Coordinating multi-step work
DAG engines, task queues, workflows, sagas
Not every project needs all 13. A simple CRUD app might only have CONVERSE, PERSIST, PRESENT, PROTECT, and BUILD. A complex AI platform might use all 13 plus custom ones.
IV. The Clustering Workflow
Step 1: Inventory First
You can't cluster what you haven't seen. Before clustering, you need a component inventory — a list of every significant component with its location, approximate LOC, and current status.
If you've already run the repo-status skill (Phase 2), use that inventory. Otherwise, walk the filesystem yourself:
# Get top-level shapels -la project-root/
# Recursively explore significant directories
find project-root/src -type f | head -50
find project-root/backend -type d
# Count LOC per directory
find project-root/backend -name "*.go" | xargs wc -l | tail -1
Step 2: Assign Verbs
For each significant component in your inventory, ask: "What verb describes what this does?"
Rules of thumb:
Most components map to one verb. If you can't decide, pick the one that best describes the component's primary purpose.
Some components legitimately serve two verbs (cross-cluster). This is fine — note both.
If a component maps to three or more verbs, it's probably doing too much. Flag it as an architectural concern.
If a component doesn't fit any verb, it might be dead code, or you might need a new verb (see Section V).
Step 3: Build Cluster Tables
For each verb that has components, create a subsection:
This table is gold for understanding coupling. Components that are cross-cluster are integration points — they're where changes in one capability can break another.
Step 5: Identify Orphans
Walk the directory tree and check: is every significant directory represented in at least one cluster?
Orphan directories — significant code that doesn't map to any cluster — signal one of three things:
Dead code that should be removed.
An emerging capability that deserves its own verb.
A gap in your analysis that needs a second look.
Document orphans explicitly. Don't sweep them under the rug.
Step 6: Write Health Assessments
For each cluster, write a 2-line health assessment:
Health line: Overall emoji + one-sentence verdict.
Audit Notes line: Key constraints, risks, or technical details.
Be honest. A cluster with 85% test coverage and active development is ✅. A cluster with no tests and a known security gap is ⚠️. A cluster that's completely broken is ❌.
V. Creating New Verbs
The 13 starter verbs are a starting point. Your project may need verbs not on the list.
Good custom verbs are specific and immediately communicative:
Verb
Use When
TRANSLATE
i18n-heavy apps, multi-language support
SIMULATE
Apps with simulation engines, digital twins
COMPOSE
Content creation tools, editors, IDEs
GOVERN
Apps with complex compliance, policy, approval workflows
Apps where notification delivery is a core capability
Bad custom verbs are vague and don't tell you anything:
MANAGE (manage what?)
PROCESS (process what?)
HANDLE (handle what?)
DO (everything "does" something)
The test: can someone read just the verb name and guess what kinds of components belong in that cluster? If yes, it's a good verb.
VI. Common Pitfalls
1. Over-clustering. Don't create 20 clusters for a 30-file project. If a cluster has only 1-2 components, merge it with a related cluster. A good rule: 4-8 clusters for small projects, 8-15 for large ones.
2. Under-clustering. Don't dump everything into PRESENT and PERSIST. If a cluster has 25+ components, consider splitting it. PRESENT might split into PRESENT (layout) and COMPOSE (content editing).
3. Confusing location with behavior. A component in frontend/src/components/ isn't automatically PRESENT. A dashboard that displays traces belongs in OBSERVE. A form that manages API keys belongs in PROTECT. Let the behavior dictate the cluster, not the directory.
4. Ignoring tests. Tests belong in BUILD, not in the cluster of the code they test. They're infrastructure, not capabilities.
5. Forgetting infrastructure. CI/CD, Docker, and deployment configs are real capabilities. BUILD is a real cluster, not an afterthought.
6. Treating clusters as hierarchical. Clusters are flat. REASON is not "above" ACT. They're peers — different capabilities of the same system. The component tables within each cluster may have internal hierarchy, but the clusters themselves don't.
VII. Using Clusters Beyond Status Documents
Semantic clusters have applications beyond the status-template:
Architecture Decision Records (ADRs): Frame decisions by which cluster they affect. "This ADR impacts REASON and ORCHESTRATE."
Sprint planning: Assign work by cluster. "This sprint we're focused on OBSERVE and PRESENT."
Code review: Ask "which cluster does this PR touch?" A PR that modifies 4+ clusters deserves extra scrutiny.
Onboarding: Walk new team members through clusters, not directories. "Let me explain what this system can DO."
Technical debt tracking: Rate each cluster's health independently. Focus debt reduction on ⚠️ and ❌ clusters.
VIII. Quality Checklist
Before delivering your semantic cluster map, confirm:
Every significant component maps to at least one cluster
No cluster has fewer than 2 components (merge if so)
No cluster has more than 20 components (split if so)
Cross-cluster components are explicitly listed
Orphan directories are documented and explained
Each cluster has a health assessment with emoji + notes
Custom verbs (if any) pass the "can you guess the contents?" test
The cluster map covers both frontend and backend (if applicable)
LOC estimates are approximate but not fictional
The map tells a coherent story about what the system does
Output
A semantic cluster map with one subsection per verb, each containing: a component table (Component, Location, Status, LOC), a one-line health assessment, and audit notes.
A cross-cluster components table listing integration points and their dual-cluster membership.
An orphans section documenting any directories that didn't map to any cluster, with an explanation.
Examples
Scenario 1: "Walk me through what the Dojo Gateway actually does" → 9 clusters identified (CONVERSE, REASON, PROTECT, OBSERVE, PERSIST, CONNECT, BUILD, ORCHESTRATE, ACT). Cross-cluster table shows TraceLogger serves both OBSERVE and BUILD. One orphan directory found: /legacy/ — documented as dead code candidate.
Scenario 2: "Plan the refactor of the frontend" → PRESENT cluster has 25 components — flagged for splitting into PRESENT (layout) and COMPOSE (content editing). 3 components mapping to 4+ clusters flagged as architectural concerns requiring decomposition.
Edge Cases
If a component maps to three or more clusters, do not force it into one — flag it explicitly as an architectural concern and note which capabilities are bleeding into each other.
Tests always belong in BUILD, never in the cluster of the code they test — even if they live adjacent to the tested components.
Anti-Patterns
Assigning components to clusters based on directory location rather than behavior — a dashboard in frontend/components/ that displays traces belongs in OBSERVE, not PRESENT.
Creating clusters with vague verbs like MANAGE, PROCESS, or HANDLE — if someone cannot guess the contents from the verb name alone, it's not a good cluster verb.
