Research Orchestration
Parallel research agent orchestration dispatching 5-10 concurrent agents for...
Install
npx skills add a5c-ai/babysitter --skill research-orchestration -g -a claude-code -y- Repository
- a5c-ai/babysitter1.9k
- Category
- AI and agents
- Updated
- 3d ago
- License
- MIT
- Context cost
- Lightabout 200 tokens
- Runs scripts
- No, instructions only
What it does
Parallel research agent orchestration dispatching 5-10 concurrent agents for comprehensive multi-source research with synthesis and validation.
Process Flow
- Plan: Decompose query into independent sub-queries
- Dispatch: Run 5-10 research agents in parallel
- Synthesize: Merge findings, identify consensus and conflicts
- Validate: Cross-reference against codebase for accuracy
Sources
- Codebase: source files, patterns, implementations
- Documentation: README, JSDoc, inline comments
- Configuration: package.json, tsconfig, CI/CD configs
Confidence Scoring
Overall confidence is a weighted average of individual agent confidence scores, adjusted by validation results. Below 70% triggers human review.
When to Use
/research [query]slash command- Before specification creation
- When investigating unfamiliar parts of the codebase
Processes Used By
claudekit-research(primary consumer)
