Agent skill
brownfield-chat
Natural-language Q&A across the full codebase. Use for multi-module questions, "what breaks if", git history, cross-cutting queries, and anything spanning more than one file. For a single-file lookup use brownfield-query instead.
Install this agent skill to your Project
npx add-skill https://github.com/wednesday-solutions/ai-agent-skills/tree/main/skills/brownfield-chat
SKILL.md
When to use
- "What breaks if I change X?" (multi-file blast radius)
- "Who last touched auth.ts?" (git history)
- "What changed in the last 30 days?" (git diff summary)
- "Which files have no tests and high risk?" (graph filter)
- "What does tokenService do?" (summary lookup)
- "What does X import / who imports X?" (structural lookup)
- Any question spanning multiple modules or layers
When NOT to use
- Architecture rules / boundary enforcement → use brownfield-drift
- About to edit a file → use brownfield-fix first
How to answer — by question type
Answer ALL questions by reading pre-built files with the Read tool.
Do NOT call wednesday-skills chat via Bash — read the files directly to avoid opening a terminal.
Summary / "what does X do?"
Read .wednesday/codebase/summaries.json→ find the file entry- If not in summaries,
Read .wednesday/codebase/MASTER.md→ search for the file section - Report the summary + risk score + blast radius count
Structural / "what does X import / who imports X?"
Read .wednesday/codebase/dep-graph.json→nodes["<file>"]- Report
imports[],importedBy[],exports[],riskScore,isEntryPoint
Blast radius / "what breaks if I change X?"
Read .wednesday/codebase/dep-graph.json- BFS over
importedBystarting from the target file (depth ≤ 5) - Report: direct dependents, transitive count, any cross-language hits
Graph filter / "which files have risk > 80?"
Read .wednesday/codebase/dep-graph.json→ iteratenodes- Filter by the requested criteria (riskScore, lang, isEntryPoint, etc.)
- Return ranked list
Git history / "who wrote X / what changed recently?"
Bash(git log --follow --oneline -20 -- <file>)for file historyBash(git log --since="30 days ago" --oneline)for recent changes- Git history is the only case that needs a Bash call
Path traversal / "how does a request reach X from Y?"
Read .wednesday/codebase/dep-graph.json- BFS through
importsedges from source to target (depth ≤ 6) - Report the shortest path found
Architecture overview
Read .wednesday/codebase/MASTER.md— entry points, primary flows, danger zonesRead .wednesday/codebase/analysis/legacy-report.json— god files, circular deps
Source citation
Always end your answer with the source used:
dep-graph.json— structural answersummaries.json— cached summaryMASTER.md— architecture / danger zonegit log— history / authorshipnot-mapped— data missing, tell dev to runwednesday-skills fill-gaps --file <file>
Never
- Call
wednesday-skills chatvia Bash — read the files directly instead - Send the full dep-graph.json to any LLM — read only the relevant nodes
- Answer from Claude training knowledge about this specific codebase
- Guess when graph data is missing — "Not mapped" is the correct answer
- Load more than 20 nodes into any LLM call
- Read raw source files (*.ts, *.go, etc.) to answer structural questions
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
wednesday-design
Design and UX guidelines for Wednesday Solutions projects. Covers visual design tokens, animation patterns, component standards, accessibility, and user experience best practices for React/Next.js applications. ENFORCES use of approved component libraries only.
brownfield-gaps
Improves graph coverage for a specific file with dynamic/unannotated patterns (event emitters, dynamic require, global injection). Use when dep-graph shows gaps on a file or when chat/query returns "not mapped". Not for querying — only for improving coverage.
onboard-dev-agent
Use when a dev is new to the project, asks for an overview, or wants to get oriented quickly. Fills graph gaps then runs a guided codebase interview.
pr-review-agent
Full PR review orchestrator. Use when asked to review a PR or check if it is ready to merge. Runs blast radius + drift check on changed files, then hands off to pr-review for comment triage and fixes. Do NOT use pr-review directly for full reviews — use this.
module-audit-agent
Use when asked to audit a module, check its health, or assess whether it should be refactored. Runs structural query, risk check, and test generation automatically.
deploy-checklist
Pre-deploy and post-deploy checklist skill. Ensures env vars, migrations, CI, rollback plan, smoke tests, and monitoring are verified before and after every deployment.
Didn't find tool you were looking for?