Claude Research

Numerically stress-test a self-authored mathematical claim over its parameter space to seek counterexamples or characterize violations. Use when checking monotonicity, thresholds, inequalities, comparative statics, or limits computationally. For algebraic proof or Lean formalization, use $symbolic-check or $lean-check.

Systematic reviewer response workflow: parse comments, classify by severity, develop response strategy, write structured rebuttal. Use when asked to 'write rebuttal', 'respond to reviewers', 'draft review response', or 'handle R&R'.

Deliver a pre-flight duplicate check before creating new skills or agents. Use when the user requests a pre-flight duplicate check before creating new skills or agents.

Query and update recorded planning and research portfolio state across tasks, topics, papers, outputs, submissions, venues, people, institutions, and deadlines. Use when asking what is already recorded, including institution- or venue-associated topics, or when doing daily planning and weekly review. For venue suitability recommendations, inspect recorded state first and then use the venue-recommendation workflow.

Recover the structure, intent, and lineage of old code, data, or analysis files. Use when inherited or dormant research code must be understood before it is changed. Not for a quality review of already-understood code.

Review user-facing documentation for accuracy, consistency, and completeness across private, public, nested repos, and the user manual. Use when docs feel stale, after major changes, or before sharing. (Replaces `repo-doc-audit`)

Manage old recordings — find large files, archive old meetings, delete processed originals. Use when the user says "clean up recordings", "how much space are meetings using", "delete old recordings", "archive meetings", "manage meeting storage", or asks about disk space from minutes.

Post-meeting debrief — analyzes what happened, compares outcomes to your prep intentions, tracks decision evolution. Use when the user says "debrief", "what just happened in that meeting", "what did we decide", "debrief that call", "post-meeting", "what changed", or right after stopping a recording.

Deliver a fast pre-commit safety scan: file size, anonymity (author / affiliation strings in tex/bib), hardcoded secrets, and invisible-Unicode carriers. Use when the user requests a fast pre-commit safety scan: file size, anonymity (author / affiliation strings in tex/bib), hardcoded secrets, and invisible-Unicode carriers. Triggers: 'audit before commit', 'check before push', 'pre-commit scan', 'safety check'.

Retarget an active manuscript to a different journal by reconciling venue identity, files, bibliography, citations, and submission-facing metadata. Use when a new target journal has already been chosen and the paper package must change. Not for choosing a venue; use the venue-fit workflow.

Reconcile the current project's Claude and Codex permission surfaces with the canonical global policy. Use when project permissions are missing, stale, or inconsistent after setup or migration. Not for syncing skills, hooks, or agents; use $sync-ai-infra.

Identify recurring weak arguments, unsupported assumptions, and vulnerable inference patterns across a literature corpus. Use when stress-testing a body of work rather than reviewing one manuscript. For one paper's argument, use the appropriate paper-review workflow.

Compile literature, meeting notes, session logs, and code findings into a structured per-project knowledge wiki. Use when project evidence is scattered across source records and needs synthesis into durable articles. Not for promoting existing articles into the global vault wiki; use $wiki-grow.

Design empirical studies through power analysis, pre-analysis planning, QSF parsing, and survey architecture. Use when specifying sampling, measurement, treatment, or analysis before data collection. Not for causal identification alone; use $causal-design.

Explore a research question through several independent analytical perspectives and synthesize their agreements and disagreements. Use when one line of reasoning is insufficient and distinct viewpoints should be preserved.

Build a traceability manifest linking analysis scripts to inputs, outputs, and manuscript figures or tables. Use when documenting a computational pipeline or locating the source of a reported artefact. Not for verifying that paper claims match code outputs; use the code-paper auditor.

Improve the exposition and readability of a mathematical proof already verified as correct without changing its mathematics. Use when polishing a lemma, theorem, proof, or appendix after correctness checks. Not for verifying the proof; use $verify-math.

Map claims in a literature to independent replications, robustness checks, failures, and unresolved evidence gaps. Use when assessing the empirical reliability of a body of findings rather than reproducing one project's code. For package rerunnability, use $replication-package.

Autonomously diagnose a codebase, apply minimal fixes, and rerun tests until they pass or a real blocker is reached. Use when the user explicitly requests an iterative fix-until-green loop across Python, R, Julia, or HPC workflows.

Bootstrap a lightweight project with minimal guidance, context, and repository structure. Use when a small non-research project needs durable AI collaboration without the full research or course scaffold. Not for formal research projects; use $init-project-research.

Compile all LaTeX projects and report cross-project build consistency. Use when checking whether a collection of papers builds cleanly. Not for rendered visual inspection after a clean build; use $latex-polish.

Turn external referee correspondence or internal pre-submission feedback into a provenance-safe, DAG-validated revision master plan with atomic tasks, dependencies, critical path, and execution blocks. Use when feedback exists but must be converted into an approved implementation sequence. Not for drafting the reviewer response itself; use $review-response.

Deduplicate and reconcile multiple completed review reports into one prioritised revision plan with conflicts and dependencies made explicit. Use when parallel reviewers have returned findings that need a single action sequence. Not for running the reviews; use $review-cluster.

Update a project's own CLAUDE.md, README.md, or docs/ to reflect current state. Use when the user asks to update a project's own CLAUDE.md, README.md, or docs/ to reflect current state.

Bootstrap a university course or module folder with the canonical teaching structure, guidance, and starter files. Use when creating or migrating a course workspace. Not for checking an existing course without modifying it; use $audit-project-course.

Assemble, anonymize, validate, or audit a research replication package. Use when preparing code and permitted data for reviewer or public release. Not for auditing code quality alone; use $code-suite or $replication-audit as appropriate.

Add or repair an editable shared Python-package dependency across consuming projects, including clone-local configuration and verification. Use when multiple repositories must import the same local package source. Not for one project's ordinary environment setup; use $python-env.

Replicate a quantitative analysis in a second language (R↔Python↔Stata↔Julia) and compare outputs for implementation errors. Use when an existing empirical result needs independent cross-language verification. Not for reviewing one implementation in place; use $code-suite.

Check a LaTeX coursework submission against the requirements in a supplied PDF assessment brief. Use when verifying format, required sections, word limits, or deliverables before submission. Not for general prose proofreading; use $proofread.

Diagnose and recover Mac Mini remote-access failures across Tailscale, mosh, resolver state, and public relays. Use when mosh hangs, the headless Mini disappears after a Tailscale change, or RustDesk/VNC/AnyDesk cannot reach its relay. Not for ordinary machine synchronization.

Scaffold, execute, analyse, and publish computational research experiments through a reproducible staged workflow. Use when a research question requires simulations or computational sweeps rather than a one-off script.

Create or migrate project-level agents, repeatable project workflows, and planning state from one client-neutral contract, then render repository-scoped adapters for both Claude Code and Codex. Use when a research project needs role separation, project commands, formal phase tracking, or conversion from an existing Claude-only .claude/agents and .claude/commands setup.

Audit a paper and submission package against current official venue requirements for its venue, track, article type, cycle, and stage. Use when checking templates, limits, anonymity, declarations, or required files before submission or inside $pre-submission-report. Not for venue recommendations.