KevinZai ccc-research
Load ONE skill. Get the entire research domain. From multi-source deep dives to spec interviews to cross-model validation.
Sub-Skills
| # | Skill | Command | Description |
|---|---|---|---|
| 1 | deep-research | /deep-research |
Multi-source deep research with parallel agents, citation tracking, and synthesis |
| 2 | spec-interview | /spec-interview |
5-7 question interview to create detailed specifications before coding |
| 3 | cross-model-review | /cross-model-review |
Review code/decisions using multiple AI models for diverse perspectives |
| 4 | literature-review | /literature-review |
Academic/technical literature review with source evaluation |
| 5 | competitive-analysis | /competitive-analysis |
Analyze competing products/tools/libraries for feature comparison |
| 6 | data-ingestion | /data-ingestion |
Ingest and summarize large documents, codebases, or datasets |
| 7 | trend-analysis | /trend-analysis |
Analyze trends in technology, markets, or usage patterns |
| 8 | gemini-fallback | /gemini-fallback |
Use Gemini's 1M context window for tasks that exceed Claude's context |
How To Use
Step 1: Tell me what you need to research. I'll route to the right specialist.
Step 2: If the task involves multiple sources or models, I'll confirm scope, depth, and output format before proceeding.
Step 3: The specialist skill handles the work. You get structured research output without loading 8 separate skills.
Routing Matrix
| Your Intent | Route To | Don't Confuse With |
|---|---|---|
| "Research this topic deeply" / "Find everything about X" | deep-research |
literature-review (academic focus, not general) |
| "Interview me to write a spec" / "Help me define requirements" | spec-interview |
deep-research (research, not requirements gathering) |
| "Get a second opinion from another model" / "Cross-validate" | cross-model-review |
deep-research (source diversity, not model diversity) |
| "Review the academic literature on X" / "What does the research say?" | literature-review |
deep-research (broader scope, less rigorous sourcing) |
| "Compare these tools" / "What's the competition doing?" | competitive-analysis |
trend-analysis (temporal patterns, not feature comparison) |
| "Summarize this codebase" / "Ingest this dataset" | data-ingestion |
deep-research (research synthesizes, ingestion summarizes) |
| "What's trending in X?" / "How is Y changing over time?" | trend-analysis |
competitive-analysis (snapshots, not trajectories) |
| "This is too big for Claude's context" / "Need Gemini for this" | gemini-fallback |
data-ingestion (processing strategy, not model switching) |
Campaign Templates
New Project Research
spec-interview-> gather requirements through structured questionscompetitive-analysis-> evaluate existing solutions in the spacedeep-research-> fill knowledge gaps identified during spec and analysisliterature-review-> find academic/technical foundations if applicable- Deliver: comprehensive spec document with competitive landscape and research backing
Technology Evaluation
deep-research-> understand the technology landscapecompetitive-analysis-> compare specific tools/libraries/frameworkstrend-analysis-> identify adoption trajectories and community momentumcross-model-review-> validate conclusions against multiple AI perspectives- Deliver: technology evaluation report with recommendation and risk assessment
Large Codebase Understanding
data-ingestion-> ingest and summarize the codebase structuregemini-fallback-> use 1M context for full-codebase analysis if neededdeep-research-> research unfamiliar patterns or libraries found- Deliver: codebase architecture summary with key patterns and dependencies
Context Strategy
This CCC domain uses on-demand loading. Sub-skills have disable-model-invocation: true so they only load when explicitly invoked, keeping your context lean.