research-suite

Plugin research-suite from marketplace scientific-computing-workflows in imewei/MyClaude. Commands read from the repo's manifest — 26 days ago.

Skills (17)

• _research-commons

  Internal-only resource directory for the research-spark skill stack. Do not select this skill directly for user prompts. Load it only when another research-spark skill (spark-articulator, landscape-scanner, falsifiable-claim, theory-scaffold, numerical-prototype, experiment-designer, premortem-criti

• evidence-synthesis

  Conduct systematic literature reviews (PRISMA), meta-analyses, and evidence grading (GRADE). This skill should be used when the user asks to "systematic review", "meta-analysis", "PRISMA flow diagram", "grade this evidence", "pool effect sizes", "compute I-squared / Q-statistic", "GRADE summary tabl

• experiment-designer

  Stage 7 of the research-spark pipeline. Translates the Stage 6 predicted observable into an experimental plan with an instrument capability map (3x margin rule on every dimension), pre-registered success metrics, controls, a formal power analysis, a design-of-experiments matrix, and a risk register.

• falsifiable-claim

  Stage 3 of the research-spark pipeline. Converts the Stage 1 articulation plus the Stage 2 landscape into a falsifiable claim with an explicit kill criterion, shaped by a Fermi feasibility screen, the Heilmeier catechism, and a Reviewer 2 adversarial pass. Triggers on phrases like "sharpen the hypot

• landscape-scanner

  Stage 2 of the research-spark pipeline. Conducts a structured three-layer literature scan (foundational / recent / adjacent), steelmanner each paper, builds a gap matrix mapping phenomena against methods, and runs a Reviewer 2 adversarial pass against the proposed research gap. Triggers when the use

• numerical-prototype

  Stage 6 of the research-spark pipeline. Converts the Stage 4-5 formalism into a running numerical prototype (JAX or Julia, chosen per physical system) and produces a concrete predicted observable, after passing three required validation passes (analytic-limit recovery, synthetic benchmark, convergen

• premortem-critique

  Stage 8 of the research-spark pipeline. Writes a failure narrative for the full plan, clusters failure modes by root cause, identifies the cheapest early signal for each cluster, and loops back to insert those signals as concrete milestones in the Stage 7 plan. Also generates a one-page summary for

• research-hub

  Top-level router for scientific research workflows. Use for: reviewing or critiquing a manuscript/paper/peer review/referee report/journal submission/novelty assessment/statistical claims/data integrity; refining a rough research idea through structured stages (research spark, falsifiable claim, the

• research-methodology

  Design-phase guide for scientific research — experimental design, hypothesis formulation, power analysis, sample-size justification, ablation planning, and statistical-test selection *before* data collection. This skill should be used when the user asks to "run a power analysis", "justify this sampl

• research-paper-implementation

  Translate published papers into working code by extracting architecture specs, hyperparameters, and training procedures from methods + appendix + supplementary material. This skill should be used when the user asks to "reproduce this paper", "implement the architecture from X", "extract hyperparamet

• research-practice

  Meta-orchestrator for the research lifecycle. Routes to specialized skills for *designing* experiments (power analysis, DoE, hypothesis planning), *evaluating* existing work (CONSORT/STROBE/PRISMA, red-flag detection), *reproducing* published papers, *writing* manuscripts (IMRaD, reports), and *synt

• research-quality-assessment

  Evaluate scientific research quality against CONSORT/STROBE/PRISMA/MOOSE reporting guidelines, score methodology and statistical rigor, and detect red flags (p-hacking, HARKing, selective reporting, circular analysis). This skill should be used when the user asks to "assess this manuscript", "audit

• research-spark

  Orchestrator for a research-refinement pipeline. A five-stage core (spark, landscape, claim, theory) turns a rough idea into a testable, fundable research proposal; three further stages (numerical prototype, experiment design, premortem) are an optional extension toward execution, not required for t

• scientific-communication

  Structure scientific arguments and write technical reports (IMRaD), abstracts, posters, and grant-style one-pagers. This skill should be used when the user asks to "structure the Introduction", "write the abstract", "IMRaD outline", "turn this into a paper", "draft the methods section", "make a one-

• scientific-review

  Use this skill whenever the user wants a formal peer review of a scientific manuscript, journal-style referee comments, reviewer recommendations, or structured review feedback on someone else's paper. Trigger on phrases like 'review this paper', 'review this manuscript', 'peer review', 'referee repo

• spark-articulator

  Stage 1 of the research-spark pipeline. Compresses a rough research idea into a 3-5 sentence articulation that names what the idea is, what is surprising about it, and what observation would confirm it. Triggers on early-stage phrases like "I had this thought about...", "rough idea:", "what if we co

• theory-scaffold

  Stages 4 and 5 of the research-spark pipeline (merged because derivation and formalization interleave in practice). Derives a theoretical framework from first principles, formalizes it in LaTeX, checks known limits, and identifies governing dimensionless groups. Triggers on phrases like "derive the

Slash commands (3)

• /research-suite:lit-review

  Structured literature review — topic scan, claim extraction, evidence synthesis, gap identification, and output as summary, table, or annotated bibliography.

• /research-suite:paper-implement

  Implement a paper's core method in JAX or Julia — parse equations, scaffold the algorithm, wire up a minimal experiment, and run smoke validation. For full claim extraction and tolerance-based replication against reported numbers, use /replicate.

• /research-suite:replicate

  End-to-end replication pipeline — fetch paper, extract claims, implement in JAX or Julia, validate outputs against reported numbers within tolerance.

Subagents (2)

• research-expert

  Use this agent for research methodology, evidence synthesis, and scientific figures, covering literature reviews, experimental design and power analysis, systematic reviews (PRISMA) and meta-analysis (GRADE), statistical-rigor audits (p-hacking, HARKing, effect sizes), IMRaD structuring, and publica

• research-spark-orchestrator

  Use this agent to drive the research-spark pipeline, turning the user's rough research idea into a testable, fundable proposal through a five-stage core (articulation, landscape scan, falsifiable claim, theory derivation) plus an optional extension (numerical prototype, experiment design, premortem)

Hooks (4)

• SubagentStop

  python3 "${CLAUDE_PLUGIN_ROOT}/hooks/subagent_stop.py"

• PostToolUse

  python3 "${CLAUDE_PLUGIN_ROOT}/hooks/post_tool_use.py"

• TaskCompleted

  python3 "${CLAUDE_PLUGIN_ROOT}/hooks/task_completed.py"

• SessionStart

  python3 "${CLAUDE_PLUGIN_ROOT}/hooks/session_start.py"

Similar plugins