repowise-intelligence
Skill12 starsChanged 29 days agoMITView on GitHub
Implements codebase intelligence layers similar to repowise-dev/repowise. Provides dependency graph, git history, auto-generated documentation, and architectural decisions intelligence for AI agents.
In this repository22 other files it gives its agents.
- AGENTS.mdAGENTS.md
- docs/agents.mdAGENTS.md
- agent/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- router/CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- .cursor/rules/verification.mdcCursor rule
- agentic-agileSkill
- agentic-portfolioSkill
- client-onboardingSkill
- cooldown-resumeSkill
- drillSkill
- ecc-harness-patternsSkill
- fabric-patternsSkill
- graphiti-temporalSkill
- managed-agents-dreamsSkill
- model-catalogue-updateSkill
- platform-setupSkill
- risky-module-reviewSkill
- seo-audit-reportSkill
- user-researchSkill
- video-contextSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 27
When an agent needs to understand "why" code was built a certain way
- Distinctive: line 28
Before making significant changes to understand impact and relationships
- Distinctive: line 29
When trying to reduce token usage by getting targeted context instead of reading full files
- Distinctive: line 30
When implementing features that benefit from understanding code ownership and change patterns
- Distinctive: line 31
When needing to answer complex questions about the codebase efficiently
- Distinctive: line 110
Initial Analysis: Run comprehensive analysis to build all four intelligence layers
- Distinctive: line 112
Efficient Storage: Use efficient data structures (JSON, SQLite) for quick retrieval
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Skill: repowise-intelligence
Purpose
Provides codebase intelligence layers inspired by repowise (https://github.com/repowise-dev/repowise). Enables AI agents to understand the repository at a deeper level through:
- Dependency graph intelligence (file and symbol relationships)
- Git intelligence (history, ownership, co-changes)
- Documentation intelligence (auto-generated docs with freshness scoring)
- Decision intelligence (architectural decisions from git history)
When to Use
- When an agent needs to understand "why" code was built a certain way
- Before making significant changes to understand impact and relationships
- When trying to reduce token usage by getting targeted context instead of reading full files
- When implementing features that benefit from understanding code ownership and change patterns
- When needing to answer complex questions about the codebase efficiently
Directory Structure
This skill creates and manages:
.claude/skills/repowise-intelligence/
intelligence/ ← Intelligence data storage
dependency_graph.json ← File and symbol dependency graph
git_history.json ← Git history analysis (hotspots, ownership, co-changes)
documentation/ ← Auto-generated documentation for modules/files
decisions.json ← Architectural decisions extracted from git history
INDEX.md ← Overview of available intelligenceIntelligence Layers
1. Graph Intelligence (Dependency Graph)
- Parses files to build file-level and symbol-level dependency graphs
- Handles import aliases, barrel re-exports, namespace imports
- Tracks heritage (extends, implements, etc.)
- Uses community detection to find logical modules
- Calculates centrality measures (PageRank, betweenness) to identify important code
2. Git Intelligence
- Analyzes git history for:
- Hotspot files (high churn × high complexity)
- Ownership percentages per contributor
- Co-change pairs (files that change together without import links)
- Significant commit messages explaining evolution
3. Documentation Intelligence
- Auto-generated documentation for modules and files
- Rebuilt incrementally on changes
- Includes coverage tracking and freshness scoring
- Supports semantic search
4. Decision Intelligence
- Extracts architectural decisions from:
- Git history (commit messages with WHY/DECISION/TRADEOFF patterns)
- Inline markers in code
- Explicit decision records
- Links decisions to the code they govern
- Tracks decision staleness as code evolves
MCP Tools Provided
This skill provides the following tools (to be implemented in agent/tools.py):
get_overview(): Architecture summary, module map, entry points, git healthget_answer(question): One-call RAG over documentation with confidence gatingget_context(targets, include?): Workhorse tool for docs, symbols, ownership, freshnesssearch_codebase(query): Semantic search over documentationget_risk(targets?, changed_files?): Hotspot scores, dependents, co-change pairsget_why(target): Get architectural decisions related to targetsget_decision_flownodes(): Extract decision-linked flow nodes
Tool Parameters
targets: Can be files, symbols, or modules (supports wildcards and globs)include: Options for get_context: "source", "callers", "callees", "metrics", "community"- In multi-repo mode:
repoparameter to target specific repository
Example Usage
# Get overview of the codebase
overview = get_overview()
# Understand why auth works a certain way
why_auth = get_why(["auth/*.py", "key_store.py"])
# Get context for modifying a specific function
context = get_context(["agent.tools:build_planning_prompt"],
include=["source", "callers", "callees"])
# Get an answer to a specific question
answer = get_answer("How does the agent planning system work?")
# Find risky areas before making changes
risk = get_risk(changed_files=["proxy.py"])Implementation Approach
- Initial Analysis: Run comprehensive analysis to build all four intelligence layers
- Incremental Updates: On each change, update only affected intelligence components
- Efficient Storage: Use efficient data structures (JSON, SQLite) for quick retrieval
- Integration: Hook into file change events to keep intelligence current
- Agent Access: Provide tools that agents can call to access intelligence
Related Skills
dependency-audit: Reviews and validates dependencies (complements graph intelligence)docs-sync: Keeps documentation current (complements documentation intelligence)graphify: Knowledge-graph token optimisation (complements graph intelligence)repo-memory-updater: Keeps CLAUDE.md and .claude/state/ in sync
Acceptance Checks
- Intelligence directory structure exists
- Tools for all six intelligence functions are implemented
- Dependency graph intelligence layer functional
- Git intelligence layer functional
- Documentation intelligence layer functional
- Decision intelligence layer functional
- Tools demonstrate significant token reduction vs naive approaches
- Integration with agent system demonstrated
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

