From do-stuff-helper
This skill should be used when the user wants to "research a topic", "look up information about X", "find out about X", or when another skill needs to conduct multi-angle web research and return a structured summary. Performs broad web searches, identifies knowledge gaps, offers deep research handoff, and synthesizes findings into a persistent research artifact.
How this skill is triggered — by the user, by Claude, or both
Slash command
/do-stuff-helper:researchThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Conduct multi-angle web research on a topic, identify gaps, offer deep research handoff, and save a structured summary to the activity's research archive.
Conduct multi-angle web research on a topic, identify gaps, offer deep research handoff, and save a structured summary to the activity's research archive.
Complete each step in order.
Once you have the main question, identify 3–5 specific sub-questions that together would answer it comprehensively. Present these to the user briefly so they can confirm or adjust the scope.
Conduct 5–8 web searches using WebSearch, targeting different angles:
For each search, use WebFetch to read the 2–3 most relevant results in full.
Parallelization: Use the Task tool with subagent_type=general-purpose to run up to 3 parallel search agents, each covering a different angle. Give each agent a clear research angle and instruct it to return structured findings with source URLs.
Evaluate which sub-questions from Step 1 are:
Identify critical unknowns that web search alone cannot resolve.
If significant gaps exist (thin, conflicting, or unanswered sub-questions on important topics), tell the user:
"I can't run deep research from Claude Code, but if you have a deep research report on [suggested topic/query] — or can run one — you can paste the content or point me to a saved file. Otherwise, I'll proceed with what I have and note the gaps."
If the user provides deep research results:
<activity-dir>/docs/research/ with a descriptive filenameIf no significant gaps exist, skip this step entirely.
Compile all results (web search + any deep research) into a structured summary:
# Research Summary: <Topic>
## Key Findings
- <finding with source citation>
## Expert Perspectives
- <perspective with source>
## Risks & Pitfalls
- <risk with source>
## Comparative Analysis
- <comparison if applicable>
## Open Questions
- <unresolved questions, gaps noted>
## Sources
- [Source title](URL) — brief relevance note
Source quality: Prefer authoritative sources (official docs, academic papers, established industry publications) over SEO content farms. Note source quality where relevant.
docs/research/ directory if no activity applies.<activity-dir>/docs/research/research-<topic-slug>.mddo-stuff-helper:research when they need domain knowledge.npx claudepluginhub ryanismert/do-stuff-helper --plugin do-stuff-helperConducts deep parallel research on a topic using web searches, saves cited markdown files with validated URLs, and produces a synthesis summary.
Conducts deep research on any topic with multi-agent source verification, interactive focus selection, and structured report generation. Supports multiple languages and session management.
Runs deep research on topics via web searches, data collection, source verification, synthesis, and structured reports saved to 02-research/. Use for reports, blogs, or exploring new domains beyond quick answers.