Agent skill
researching
Comprehensive research with cited sources. Use for complex research that should be verified and persist.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/devops/researching
SKILL.md
Researching
Produce well-sourced research artifacts with verified citations.
When to Use
- Complex research requiring multiple sub-queries
- Need documented research that persists beyond conversation
- Topic requires source verification and confidence calibration
- High-stakes decisions that need cited evidence
Don't use for: Quick lookups. Just search directly.
Artifact Requirements
- Minimum 3 independent evidence bundles per research topic
- Each bundle must cite primary sources with URLs
- Source diversity required: official docs, peer-reviewed, reputable secondary
- Academic citations must be validated with
bibval
Quality Rubric
| Criterion | Requirement |
|---|---|
| Citations | Every claim has inline citation with URL |
| Coverage | Key perspectives included (not just first result) |
| Recency | Sources current (≤2 years for APIs/tech, flexible for fundamentals) |
| Confidence | Calibrated honestly; uncertainties stated explicitly |
| Conflicts | Disagreements between sources noted, not hidden |
Source Credibility Hierarchy
Weight findings by source authority:
- Official documentation - canonical, highest weight
- Peer-reviewed papers - validated, high weight
- Reputable blogs/talks - expert authors, medium weight
- Stack Overflow (accepted) - community validated, medium weight
- Forums/unverified - low weight, note uncertainty
Output Schema
{
"topic": "string",
"status": "COMPLETE | PARTIAL",
"confidence": "HIGH | MEDIUM | LOW",
"findings": [
{
"subtopic": "string",
"content": "string with inline citations",
"sources": ["url1", "url2"]
}
],
"sources": [
{
"url": "string",
"title": "string",
"authority": "official | peer-reviewed | blog | forum"
}
],
"gaps": ["string"]
}
Constraints
- Prefer primary sources over secondary
- Cross-reference claims across multiple sources
- Flag single-source claims as lower confidence
- Run
bibvalon academic citations before including
Recording Findings
Post research artifacts to jwz for persistence and discovery:
jwz post "research:<topic>" --role alice \
-m "[alice] SYNTHESIS: <topic>
Status: COMPLETE | PARTIAL
Confidence: HIGH | MEDIUM | LOW
Findings:
- <finding 1>
- <finding 2>
Sources: <url1>, <url2>"
Discovery
jwz search "SYNTHESIS:"
jwz search "FINDING:"
jwz read "research:<topic>"
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?