From qe-framework
Automated codebase analysis for brownfield projects. Spawns 4 parallel agents to map stack, architecture, quality, and concerns.
How this skill is triggered — by the user, by Claude, or both
Slash command
/qe-framework:Qmap-codebaseThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
A skill that rapidly maps an unfamiliar or brownfield codebase by spawning 4 parallel Teammate agents. Each agent analyzes a distinct dimension of the project and writes its findings to `.qe/analysis/`. The Lead then synthesizes a single executive summary.
A skill that rapidly maps an unfamiliar or brownfield codebase by spawning 4 parallel Teammate agents. Each agent analyzes a distinct dimension of the project and writes its findings to .qe/analysis/. The Lead then synthesizes a single executive summary.
.qe/analysis/ where every other QE skill expects them./Qinit first..qe/analysis/ already contains output files (stack-profile.md, architecture-map.md, quality-report.md, concerns-report.md). If any exist, ask the user before overwriting. In Utopia mode, overwrite without asking.Launch all 4 agents in a single Agent tool call block so they run concurrently.
Goal: Identify the technology stack, language versions, and dependency landscape. Instructions:
package.json, requirements.txt, Pipfile, pyproject.toml, go.mod, Gemfile, Cargo.toml, pom.xml, build.gradle, *.csproj, composer.json, Makefile, Dockerfile, docker-compose.*.qe/analysis/stack-profile.mdOutput format:
# Stack Profile
## Language & Runtime
## Frameworks
## Key Dependencies (top 15)
## Dev Tooling
## Build & Package
Goal: Map the high-level architecture, module boundaries, and data flow. Instructions:
main.*, index.*, app.*, server.*, src/.qe/analysis/architecture-map.mdOutput format:
# Architecture Map
## Directory Structure (depth 2)
## Entry Points
## Module Boundaries
## Data Flow (inferred)
## Key Patterns (MVC, microservices, monorepo, etc.)
Goal: Assess testing, linting, CI/CD, and code health indicators. Instructions:
jest.config.*, vitest.config.*, pytest.ini, .eslintrc*, .prettierrc*, tsconfig.json, .github/workflows/*, .gitlab-ci.yml, Jenkinsfile, .circleci/*, codecov.yml, sonar-project.properties**/*.test.*, **/*.spec.*, **/*_test.*, test/**, tests/**, __tests__/**.qe/analysis/quality-report.mdOutput format:
# Quality Report
## Test Infrastructure
- Framework:
- Test file count:
- Test directories:
## Linting & Formatting
## CI/CD Pipeline
## Type Safety
## Coverage Configuration
Goal: Surface technical debt, security risks, and maintenance concerns. Instructions:
TODO, FIXME, HACK, XXX, DEPRECATED, @deprecated (count and sample top 10)console.log, debugger, binding.pry, import pdb (debug artifacts).env.example (env var management), .nvmrc/.node-version/.tool-versions (version pinning)eval(, dangerouslySetInnerHTML, innerHTML =, raw SQL concatenation.qe/analysis/concerns-report.mdOutput format:
# Concerns Report
## Technical Debt
- TODO/FIXME count:
- Top 10 samples:
## Debug Artifacts
## Dependency Health
- Lock file last updated:
- Version pinning:
## Security Signals
## Recommendations (top 5)
After all 4 agents complete, the Lead reads all output files and creates:
.qe/analysis/codebase-summary.md -- a one-page executive summary:
# Codebase Summary
Generated: {date}
## Overview
One paragraph: what this project is, primary language/framework, rough size.
## Stack Snapshot
Key technologies in a compact table.
## Architecture at a Glance
2-3 sentence description of the architecture pattern and module layout.
## Health Score
Simple assessment: Testing (Good/Fair/Poor), CI (Yes/No), Linting (Yes/No), Type Safety (Yes/No).
## Top Concerns
Bulleted list of the 3-5 most important findings from the Concerns Report.
## Recommended Next Steps
What the developer should investigate or fix first.
Print the executive summary to the user with key findings highlighted. If any dimension reported critical issues (no tests, no CI, security signals), call them out explicitly.
.qe/analysis/After displaying the summary, suggest next steps:
/Qsystematic-debugging or manual review/Qplan for phased onboarding/Qtest-driven-development to add coverage/Qrefresh to keep analysis current as work progressesnpx claudepluginhub inho-team/qe-framework --plugin qe-frameworkCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.