cc-toolkit

Use when creating, listing, searching, or deleting macOS Calendar events — scheduling meetings or appointments, checking today's or this week's agenda, or querying date ranges via the bundled EventKit CLI. Requires macOS with Calendar access.

Use when capturing thoughts or tasks into the GTD inbox, or when listing, counting, removing, or clearing inbox items.

Use when deciding what to work on next — calendar-aware task selection, time-blocked agenda generation, and agent task dispatch for the GTD Engage step.

Use when listing the whole GTD system read-only — every open project with its actions plus standalone actions not linked to any project.

Use when processing, triaging, or clarifying GTD inbox items into projects and context-list actions following the GTD clarify and organize workflow.

Use when reviewing GTD project status, adding actions or agent tasks to projects, completing or rescheduling actions, or completing projects tracked in macOS Reminders.

Use when creating, completing, listing, or deleting macOS Reminders — setting due dates, priorities, notes, or recurrence, checking pending or overdue tasks, or managing reminder lists via the bundled EventKit CLI. Requires macOS with Reminders access.

Use when creating or reviewing scientific figures for publication, including journal dimensions, resolution, file formats, typography, line weights, colorblind-safe palettes, and Nature, Science, ACS, or RSC requirements.

Use when creating or reviewing scientific presentations, conference talks, lab meetings, journal clubs, or STEM slides involving content density, visual evidence, equations, figures, typography, or cognitive load.

Use when the user requests step-based file naming and folder organization for a new feature, including sequential numbered prefixes, sub-step notation, and plan tracking across scripts, outputs, tests, or documentation.

This skill should be used when the user asks to "discuss with codex", "debate this with codex", "get codex's take", "deliberate with codex", or wants Claude and the Codex CLI to reach a reasoned conclusion on a question or design through autonomous turn-by-turn adversarial discussion. Claude holds a genuine position while Codex acts as an always-adversarial read-only critic; the loop runs to convergence or a round cap, then a conclusion is saved and presented.

This skill should be used when the user asks to "create cycle", "create the task loop", "generate the task-loop instructions", "write the per-task playbook", "scaffold task-loop", or to turn a task-loop proposal into the per-project cycle a worker follows. It renders docs/task-loop/task-loop.md — this project's tailored step-by-step worker cycle, its general rules, and its parameters (the cycle-worker agent follows this file) — from auto-detected and interviewed specifics, and scaffolds docs/task-loop/directions.md and a .gitignore entry.

This skill should be used when the user asks to "run cycle", "run the task loop", "start the orchestrator", "drive the autonomous loop", or to begin autonomous, orchestrated execution of a task-loop project. It runs the orchestrator under a fixed-interval loop; each tick honors human steering, merges ready PRs, diagnoses and re-attacks any failed/blocked/stuck attempt with a materially-different AIM-aligned strategy (it never gives up short of the goal or the time bound), reconciles the roadmap, and dispatches the next work. Idempotent — every tick re-derives state from the task DB + GitHub + directions.md.

Use when running task-loop in Codex, starting the Codex controller, dispatching task_loop_cycle_worker, processing task-loop task DB state, or executing a manual task-loop orchestration pass after setup, specify-aims, and create-cycle.

Use when a Codex task-loop project has docs/task-loop/proposal.md and needs create-cycle scaffolding, generated task-loop instructions, docs/task-loop/task-loop.md, directions.md, logs/.gitkeep, or goal-rubric scratch ignore setup.

Use when setting up task-loop in Codex, checking existing task-loop setup, configuring Supabase, saving credentials, registering a repo, syncing the Codex cycle-worker agent, checking prerequisites, running preflight, performing a setup smoke test, or preparing to use task-loop specify-aims, create-cycle, or run-cycle.

Use when starting a task-loop project in Codex, defining project aims and stages, initializing docs/task-loop/proposal.md, re-aiming a proposal before a run starts, or planning task-loop milestones and success criteria.

This skill should be used when the user asks to "set up task-loop", "configure the Supabase backend", "connect the task-loop database", "save Supabase credentials", "onboard a repo to task-loop", "check existing task-loop setup", "check task-loop prerequisites", "verify task-loop dependencies", "enable agent teams", "preflight", or otherwise prepare a machine/repo to run the task-loop harness. It checks existing machine/repo setup first, then walks through missing Supabase, credential, repo-registration, Agent Teams, and required-skill setup only as needed.

This skill should be used when the user asks to "specify aims", "initialize the proposal", "set up the task-loop proposal", "define the project aims and stages", "plan the project stages and milestones", or "start a task-loop project" — i.e. to author the living proposal that drives a task-loop autonomous development project. It produces docs/task-loop/proposal.md with three parts — specific aims & goal, an implementation plan (stages + milestones), and a living roadmap (progress) — brainstorming the aims with the user and pressure-testing the goal and stage decomposition with discuss-with-codex.

Use when a design, plan, rubric, PR, failure diagnosis, or task orchestration decision needs independent adversarial review before conclusion.

Use when the user asks to update docs, refresh documentation, fix stale README content, audit documentation quality, remove outdated prose, or bring existing documentation in line with current project truth.

Use when the user wants a measurable /goal rubric, success criteria, acceptance criteria, or completion condition, or when a /goal loop will not close because the criteria are not observable or binary.

This skill should be used when the user asks to "list all projects and actions", "show gtd overview", "show all my projects", "list everything", "what's in my gtd system", "show standalone actions", or wants a complete read-only listing of all GTD projects with their actions plus all individual actions not linked to any project.

Draft a binary (pass/fail) rubric and a ready-to-paste completion condition for a Claude Code or Codex /goal command. Use when the user wants to write or design a rubric, success criteria, acceptance criteria, or a "done"/completion condition for /goal, or when a /goal loop will not close because its criteria are not measurable.

This skill should be used when the user asks to "update the docs", "update documentation", "refresh the README", "the docs are out of date", "update X to reflect Y", "audit this doc", "clean up the documentation", or wants existing documentation brought to current truth and checked for quality. It edits existing docs in place (not issue- or PR-tracked development work), moves change history to a separate CHANGELOG, and audits the result against a documentation quality rubric.

This skill should be used when the user asks to "process inbox", "process items", "triage inbox", "clarify inbox", "organize inbox", "categorize tasks", or wants to process GTD inbox items into projects or actions following the GTD clarify/organize workflow.

This skill should be used when the user asks to "gather PR feedback", "review and fix PR issues", "run pr-feedback", "address reviewer comments", or wants to run a structured review-fix-push cycle on the current pull request. Orchestrates self-review, external review collection, issue investigation, and iterative fixing until clean.

This skill should be used when the user asks to "set up statusline", "configure statusline", "install statusline", "set up status bar", or wants to configure the Claude Code statusline with model info, git status, context/rate-limit bars, and token cost tracking.

This skill should be used when the user asks "what should I work on next?", "plan my next block", "generate an agenda", "what's next?", "dispatch agent tasks", or wants calendar-aware task selection, time-blocked agenda generation, and parallel agent task dispatch for the GTD Engage step.

This skill should be used when the user asks to "review projects", "manage projects", "check project status", "add action to project", "add agent task", "complete project", or wants to review and manage GTD projects with a guided workflow including status tracking, action management, and agent-centric task assignment.

This skill should be used when the user asks to "change model", "set default model", "switch model", "use 1M context", "enable 1M context window", "change effort level", "set thinking depth", "configure effort", "adjust auto compact", "change compact percentage", "set auto compact threshold", "configure auto compaction", "configure model settings", or wants to modify the model, effortLevel, or CLAUDE_AUTOCOMPACT_PCT_OVERRIDE settings in Claude Code.

This skill should be used when the user asks to "evaluate the project", "run a deep eval", "review this feature", "evaluate the auth module", "find issues", or wants multi-angle project evaluation with long-running critic agents. Dispatches skeptical evaluator agents that iterate through evaluation angles to find issues across code, architecture, security, UX, product, and design dimensions.

This skill should be used when the user asks to "set up discord", "configure discord bot", "install discord plugin dependencies", or wants to set up the prerequisites for the Claude Code Discord channel plugin including bun installation and shell alias configuration.

Guide for creating publication-quality scientific figures that meet major journal standards (Nature, Science, ACS, RSC). Use this skill when users ask about scientific figure formatting, journal requirements, colorblind-safe palettes, resolution/DPI, file formats, fonts, or figure dimensions for publication.

Comprehensive guide for creating effective academic presentation slides for STEM fields. Use this skill when users ask to create scientific presentations, review slide design, need guidance on slide content density, bullet point limits, text vs visuals, equation presentation, figure adaptation, or want to avoid common presentation mistakes. Applies to conference talks, lab meetings, journal clubs, and all academic presentation contexts.

Guides through standardized development workflow from brainstorming to PR merge and cleanup. This skill should be used when starting new features, bug fixes, refactoring, documentation updates, or any development work that requires issue tracking and PR workflow. Use when the user wants to follow a systematic approach to problem-solving with proper version control and code review practices.

This skill should be used when organizing work using step-based file naming and folder structure. Load this skill at the start of working on a new feature to establish sequential, numbered organization for all files (scripts, outputs, tests, documentation). The skill guides the use of numbered prefixes (01_, 02_, 03_) and sub-step notation (01_1_, 01_2_) to maintain clear workflow order and traceability.

Convert PDF documents to AI-accessible markdown format using IBM's Docling library. This skill should be used when the user needs to extract content from PDFs including text, figures, and tables in a structured markdown format. It handles scientific papers, technical documents, reports, and any PDF requiring content extraction for AI processing or analysis.

Intelligently rename research paper PDFs from generic filenames (DOIs, serial numbers) to descriptive names based on extracted titles and metadata. This skill should be used when the user has downloaded research articles with non-descriptive filenames and wants to organize them with meaningful names based on the document content.

This skill should be used when the user asks to "add to inbox", "capture a thought", "show inbox", "list inbox items", "clear inbox", "remove from inbox", or wants to manage their GTD inbox for capturing thoughts and tasks.