zy839971925-zyy/Agent--deep-research-Workflow-Skills / modular/skills/deep-research
deep-research
VerifiedUse when a scenario calls for real-world evidence, pre-build engineering/architecture research, or a linked hybrid; infer the route without requiring the user to label it. Keep the host-selected default model; infer proportionate workflow depth, honoring explicit depth requests such as Ultra.
Deep Research
This is a Modular Reasoning Workflow Skill. It preserves the governing semantics of Reasoning Workflow but is intentionally scoped to one Family. Use only the references needed for the current gap; do not bulk-load the folder.
User instructions and hard safety/authorization constraints take precedence. This Skill is a strategy and invariant layer, not a mandatory long checklist. If another Family clearly owns the next gap and that sibling Skill is installed, route to it rather than duplicating its method here.
Activate for material evidence depth, frame uncertainty, multi-source verification, causal/systemic uncertainty, competing explanations, or an explicit Deep/Max/Ultra research request. Infer real-world, engineering, or hybrid research from the scenario; users need not name the lane. The selected model remains the host default; explicit workflow depth overrides inferred depth, not model identity or permissions.
When a material ambiguity may change the research path or recommendation, first resolve what can be checked from conversation, task state, supplied materials, live host capabilities, external sources, and safe derivation. Ask only for a user-owned input that still materially changes the result; if the sibling Reasoning Core Skill is installed, use its clarification gate. Clarification is not consent. In any case, use only actually available research tools and name a fallback honestly.
Progressive references
-
research-to-report — Turn a discriminating search into a report with bounded, traceable claims.
-
research-modes-and-methods— Load when a material investigation is real-world/empirical, pre-build architecture, or hybrid. -
inquiry-and-research— Load only when the active gap matches one of: deep-research, evidence-need, research-loop, frame-discovery. -
information-target-navigation— Load when a material unknown needs a discriminating observation or the investigation needs a compact research frontier. -
evidence-and-provenance— Load only when the active gap matches one of: evidence, provenance, citation, source-independence. -
comparative-synthesis— Load only when empirical studies, model results, or benchmarks need relation-preserving comparison. -
retrieval-and-observation— Load only when the active gap matches one of: retrieval, orientation-retrieval, evidence-retrieval, observation. -
time-scenarios-and-forecasting— Load only when the active gap matches one of: forecast, time, scenario, freshness, volatility. -
multimodal-and-data— Load only when the active gap matches one of: multimodal, data, files, images. -
ultra-research— Load only for explicit Ultra depth or gaps involving fresh-frame, evidence-ecology, missingness, boundary, research-path reset, or pivotal-claim coverage.
Modular boundary
Return findings, decisions, evidence, uncertainty, and state effects to the parent task. Do not treat agreement between sibling Skills or agents as independent empirical evidence.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |