Agent skill
general-research-workflow
Core research loop for scoping, searching, analyzing, and reporting with explicit constraints and validation.
Stars
163
Forks
31
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/devops/general-research-workflow-dnyoussef-context-cascade
SKILL.md
STANDARD OPERATING PROCEDURE
Purpose
- Provide a default research workflow when no specialized skill is specified.
- Ensure constraint clarity, evidence hygiene, and confidence ceilings.
- Maintain skill-forge structure: SKILL + README + examples + references.
Trigger Conditions
- Positive: open-ended research tasks, quick investigations, or shaping a plan before deeper pipelines.
- Negative: heavy orchestration (use
deep-research-orchestrator) or prompt-only work (prompt-architect).
Guardrails
- Extract HARD / SOFT / INFERRED constraints with sources; confirm INFERRED items early.
- Two-pass refinement on deliverables: structure/coverage then epistemic validation.
- Keep outputs in English, cite evidence, and state confidence with ceilings.
Inputs
- Research question or decision to support.
- Time constraints, depth expectations, and deliverable format.
- Starting sources (if any) and exclusions.
Workflow
- Frame & Constraints: Capture the question, decision, and constraint buckets.
- Search & Collect: Gather high-signal sources; log provenance.
- Analyze: Extract claims, compare sources, and note conflicts or gaps.
- Synthesize: Organize findings into themes tied to the question and constraints.
- Validate: Check coverage vs. constraints, source quality, and confidence ceilings.
- Deliver: Provide summary, evidence table, risks, and next steps; store references/examples.
Validation & Quality Gates
- Minimum two passes completed with documented changes.
- Evidence linked to sources; conflicting evidence surfaced.
- Confidence ceilings aligned to evidence type.
- Deliverables stored with project tags.
Response Template
**Question & Constraints**
- HARD: ...
- SOFT: ...
- INFERRED (confirm): ...
**Findings**
- Claim → evidence → confidence ceiling.
**Gaps / Risks**
- ...
**Next Steps**
- ...
Confidence: 0.78 (ceiling: inference 0.70) - based on gathered evidence and validation checks.
Confidence: 0.78 (ceiling: inference 0.70) - Default setting after completing the SOP.
Didn't find tool you were looking for?