Deep research with tech stack context, parallel agents, and long-form output
From mnpx claudepluginhub molcajeteai/plugin --plugin m<research query or URL>claude-sonnet-4-6/researchPerforms adaptive deep web research on a query with configurable --depth and --strategy options. Outputs markdown report with executive summary, analysis, confidence scores, and cited sources.
/researchConducts deep multi-AI research on a topic after selecting intensity level, producing structured report with executive summary, key themes, takeaways, and sources.
/researchConducts multi-turn deep research on a codebase topic over 5 iterations, tracing code paths with citations, Mermaid diagrams, tables, and confidence ratings.
/researchGathers knowledge from trusted web sources and files, cross-references claims across multiple sources, and produces cited research documents in docs/research/. Optional skill distillation via --skill-for.
/researchRuns research phase for current or specified spec: gathers context, optional interview, parallel subagent research, synthesizes research.md, reviews, approves, and finalizes state.
/researchRuns multi-source research session on a topic across GitHub, HN, Lobsters, Reddit, arXiv, Semantic Scholar; applies TRIZ cross-domain analysis to produce a domain-appropriate report. Also supports format, resume, list, and domain options.
Research input: $ARGUMENTS
Read the research-methods skill and execute it in Deep Research mode — skip Step 1 classification and go directly to Step 2 (Detect Tech Stack).
Read: ${CLAUDE_PLUGIN_ROOT}/skills/research-methods/SKILL.md
Follow the skill's Deep Research Orchestration (Steps 2-6) exactly. The skill is the single source of truth for templates, writing style, agent prompts, synthesis format, and save flow.