zy839971925-zyy/Agent--deep-research-Workflow-Skills / integrations/deepseek-harness/skills/reasoning-workflow
reasoning-workflow
VerifiedDefault silent workflow for everyday questions, research, writing, design and engineering. Match effort to the request, resolve costly ambiguity, find specialist resources and verify the requested outcome. Explicitly selectable with “调用我的 workflow skills”.
references/domain-patterns.md
View on GitHub# Domain patterns
**Trigger:** A domain has characteristic evidence, failure, or verification patterns that materially change the generic workflow.
**Reads:** task domain, current lane, working model, effort vector.
**Updates:** domain-specific evidence/verification routes and failure checks.
**May invalidate:** generic plans that ignore domain-specific source, safety, or validation constraints.
**Exit:** domain-specific adaptations have been integrated without replacing the governing workflow.
**Related:** [inquiry and research](inquiry-and-research.md), [synthesis, execution and verification](synthesis-execution-and-verification.md).
**Return:** [root workflow](../SKILL.md) → whichever lane is active.
Use only when domain-specific methods materially change the reasoning. These are adaptations of the universal workflow, not separate mandatory procedures.
## Science and medicine
Distinguish hypothesis, mechanism, association, prediction, intervention effect, and individual outcome. Examine study design, measurement, selection, population, uncertainty, bias, corrections, and study identity. Peer review is not proof. Formal reviews need explicit eligibility, search, screening, extraction, and risk-of-bias methods appropriate to the domain.
## Technology, software, and AI
Separate documented capability, inspected implementation, configured environment, observed execution, and benchmarked performance. Resolve versions before merging evidence. For bugs and performance, inspect repo/runtime evidence before generic web commentary. Benchmarks transfer only when tasks, tools, budgets, environments, and scoring are comparable.
## Markets, business, and organizations
Combine outside-in forces with inside-out operating mechanics. Separate fundamentals, incentives, expectations, reported metrics, and market observations. Align entities, accounting periods, definitions, and point-in-time information. For decisions, identify pivotal assumptions, buffers, bottlenecks, competitive response, and signposts that would change the view.
## Policy, law, regulation, and institutions
Track jurisdiction, authority, effective dates, amendments, implementation, enforcement, and superseding text. Distinguish legal text, official interpretation, implementation practice, litigation, and stakeholder claims. “Policy exists” and “policy works as intended” are different propositions.
## History and events
Separate contemporaneous evidence from retrospective accounts. Resolve identity and chronology before causal narrative. Consider what was knowable at the time, archive/version changes, provenance of repeated stories, and alternative explanations for observed sequence.
## Products and consumer decisions
Start from the user's goals, constraints, switching/migration costs, risk tolerance, maintenance, ecosystem, total cost, and use context. Current specifications and availability may require fresh verification. User reports are useful for field behavior but need model/version/context alignment and provenance awareness.
## People, communication, negotiation
Use the human-context reference. Distinguish wording, context, plausible intent, incentives, power/relationship dynamics, and uncertainty. Do not infer stable personality, hidden motive, or diagnosis from thin evidence. In negotiation, model objectives, alternatives, constraints, information asymmetry, and likely response without pretending to know the other side's mind.
## Creative and editorial work
Research factual dependencies, audience expectations, genre conventions, references, or examples only when they materially improve the piece. The objective is the artifact, not maximal citation density. Preserve the user's intended voice and avoid laundering uncertain claims into confident prose.
## Controlled operational/process change
For changes to production systems, workflows, policies, operating procedures, or controlled artifacts, use the change-governance reference. Resolve the current baseline, affected interfaces, operator/user reality, temporary versus permanent status, readiness, rollout/rollback, monitoring, and effectiveness. “The change was implemented” and “the change is effective in operation” are separate claims.
[← Return to root workflow](../SKILL.md)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |