feature-investigation
> [IMPORTANT] Use TaskCreate to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ask user whether to skip.'
npx skills add majiayu000/claude-skill-registry --skill feature-investigation-duc01226-easy-claude --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Investigates and explains how existing features or logic work in a read-only manner. It plans via TaskCreate to break work into small tasks before reading files, reads and traces code evidence, and outputs findings with explicit file:line proof and confidence levels.
How it works
1. Setup and Scope
- Plan step-by-step tasks using TaskCreate before reading files.
- Define investigation scope and metadata in the analysis file.
2. Discovery and Reading
- Glob/Grep across backend and frontend to locate relevant files (e.g., domain entities, handlers, controllers).
- Read each targeted file to extract concrete implementation details.
3. Evidence and Validation
- For every finding, attach
file:lineproof and a confidence level (95%+, 80-94%, 60-79%, <60%). - If a claim cannot be proven by code trace, mark as inferred with low confidence.
4. Reporting
- Build a knowledge graph per file with key attributes (file path, type, dependencies, business context, etc.).
- Provide an end-to-end workflow synthesis, integration points, and business rules.
5. Output
- Return a structured review with sections directly describing what the skill does, how it works, and caveats.
Key Files
- Analysis file in
.ai/workspace/analysis/...-investigation.analysis.mdpopulated with: Metadata, Progress, Errors, Assumption Validations, Knowledge Graph, etc. - Evidence is anchored to concrete code references in
file:lineformat.
--- name: feature-investigation version: 1.1.1 description: "[Investigation] Use when the user asks to investigate, explore, understand, explain, or analyze how an existing feature or logic works. Triggers on keywords like "how does", "explain", "what is the logic", "investigate", "understand", "where is", "trace", "walk through", "show me how"." allowed-tools: Read, Grep, Glob, Task, WebFetch, WebSearch, TaskCreate --- > **[IMPORTANT]** Use `TaskCreate` to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ask user whether to skip. - `docs/project-reference/domain-entities-reference.md` — Domain entity catalog, relationships, cross-service sync (read when task involves business entities/models) > **Evidence Gate:** MANDATORY IMPORTANT MUST — every claim, finding, and recommendation requires `file:line` proof or traced evidence with confidence percentage (>80% to act, <80% must verify first). > **External Memory:** For complex or lengthy work (research, analysis, scan, review), write intermediate findings and final results to a report file in `plans/reports/` — prevents context
- Quick Summary
- Investigation Mindset (NON-NEGOTIABLE)
- ⚠️ MANDATORY: Confidence & Evidence Gate
- Prerequisites
- Anti-Hallucination Protocols
- ASSUMPTIONVALIDATIONCHECKPOINT
- EVIDENCECHAINVALIDATION
- TOOLEFFICIENCYPROTOCOL
- CONTEXTANCHORSYSTEM
- Quick Reference Checklist
- Analysis File Structure
- File Discovery Search Patterns
- File List Organization
- Knowledge Graph Entry Template
Domain Entities (HIGH PRIORITY) Commands & Queries (HIGH PRIORITY) Event Handlers (HIGH PRIORITY) Controllers (HIGH PRIORITY) Background Jobs (MEDIUM PRIORITY) Message Bus Consumers (MEDIUM PRIORITY) Services & Helpers (LOW PRIORITY) Frontend Components & Stores (HIGH PRIORITY)
What does the feature-investigation skill do?
> [IMPORTANT] Use TaskCreate to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ask user whether to skip.'
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill feature-investigation-duc01226-easy-claude --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.