Agent skill
plan-research
Research a problem space with evidence gathering and source validation.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/plan-research
SKILL.md
Plan Research
The Planner drives this workflow. Investigate a problem space by gathering evidence from the codebase, documentation, and external sources. Produce a structured research summary with cited sources.
Phase 1: Scope the Investigation
Define the boundaries before researching.
Steps
- State the research question clearly
- Identify what you need to learn and why
- Define what is in-scope and out-of-scope
- List known starting points (files, docs, APIs, prior ADRs)
Output
Produce a scoped research brief covering: research question, in/out scope, starting points.
Approval Required: No
Phase 2: Gather Evidence
Collect evidence from internal and external sources.
Internal Sources
- Search the codebase for relevant patterns, implementations, and conventions
- Read existing ADRs, architecture docs, and design notes
- Review git history for context on past decisions
- Quantify where possible (line counts, file counts, dependency counts)
External Sources
- Use
WebFetchto read official documentation, changelogs, or specifications - Use
mcp__github__search_repositoriesto explore prior art in open source - Do NOT rely solely on training knowledge for version-specific behavior, API signatures, or configuration options — verify against the source
- Cite every external source with a URL or repository reference
Evidence Standards
- Every claim must link to a specific file, line, URL, or commit
- Distinguish verified facts from inferences
- Note contradictions between sources
- Record what you searched for and didn't find (negative evidence)
Output
Produce a findings report covering: topic-by-topic findings with sources, negative evidence.
Approval Required: No
Phase 3: Synthesize
Combine findings into a structured research summary.
Steps
- Combine findings into a coherent picture
- Identify patterns, gaps, and contradictions
- State implications for the original research question
- List open questions that remain unanswered
Output
Produce a research summary covering: synthesis narrative, key insights (numbered), open questions, sources table (source | type | key finding).
⛔ CHECKPOINT
STOP. Do not proceed until human confirms the research summary is accurate and sufficient.
Approval Required: Yes
Error Handling
| Error | Recovery |
|---|---|
| Research question too broad | Narrow scope, break into sub-questions |
| External source unavailable | Note the gap, try alternative sources |
| Contradictory evidence | Document both sides, flag for human judgment |
| Insufficient evidence | State what's missing, recommend additional investigation |
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?