Explore

创建、审查和维护 agent 技能（Skill）。Use when creating/refining a Skill or deciding whether Skill、command、hook、rule、AGENTS.md guidance is the right carrier.

Conduct a retrospective on a coding session.

Investigate a question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched, docs or API facts gathered, or reading legwork delegated to a background agent.

Compact the current conversation into a handoff document for another agent to pick up.

A relentless interview to sharpen a plan or design, which also creates docs (ADR's and glossary) as we go.

A relentless interview to sharpen a plan or design.

Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes: Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue/spec asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".

Generate a HANDOFF.md context transfer document for session continuity. Use when switching sessions, handing off to another agent, or preserving complex task context before a context window reset. Trigger on "/handoff" command or "handoff:" prefix in inter-agent messages.

Generate a group-chat roast on demand. Tuned to the specific group's tone. Use with caution.

The HyperFrames composition contract — build one renderable project. Use for composition structure, the `data-*` timing attributes, `class="clip"`, tracks, sub-compositions, variables, framework-owned media playback, deterministic-render rules, and validation. Read before writing composition HTML.

Use the HyperFrames CLI development loop: init, add, catalog, capture, lint, check, snapshot, compare, grade-compare, preview, play, present, beats, keyframes, single or batch render, publish, cloud, cloudrun, feedback, lambda, doctor, browser, info, upgrade, skills, compositions, timeline, history, docs, benchmark, telemetry, transcribe, auth, tts, and remove-background. Also use when diagnosing build or render failures. validate, inspect, and layout are deprecated aliases; use check. Covers local, HeyGen-hosted cloud, AWS Lambda, and Google Cloud Run rendering.

Video and audio processing with FFmpeg. Use for format conversion, resizing, compression, audio extraction, and preparing assets for Remotion. Triggers include converting GIF to MP4, resizing video, extracting audio, compressing files, or any media transformation task.

The technique catalog: five velocity-matched SEAMS (zoom-through, INVERSE zoom-through, cut-the-curve, waterfall cut, rack-focus blur-cut) plus the two in-scene techniques — waterfall ENTRY (staggered arrival cascades for title cards / segment openers) and the nudge curve (slow-fast-slow three-phase group slides). Covers partial-travel (~12% of frame) velocity matching via mirrored power4 eases, the Z scale-sign rule, size-scaled blur (10px text / 18-20px full-frame), word-by-word staggered cuts, cascade pacing by element weight, and the 10/65/25 slide ratio. Read before authoring any transition, text-beat handoff, kinetic text entry, or group reposition. [depth, zoom, inverse-zoom, scale-sign, mirrored-zoom, rack-focus, pacing, velocity, cut-the-curve, waterfall, stagger, cascade, kinetic-text, title-card, segment-opener, nudge, slide, easing, group-motion, z-depth, motion-graphics, cinematic, transition, blur, directional-continuity]

Overlay doctrine for the embedded-captions workflow — the caption MODEL (drop / rail / embed) and the rule that captions are an OVERLAY composited on top of the film, never a reserved bottom band you shift content up to avoid. Load when adding captions/subtitles to a talking-head or launch video, when deciding whether a phrase should be dropped, ride the verbatim rail, or be promoted to a scarce embedded climax, when laying out a composition that will carry captions (do NOT reserve a keep-out band), or when centering a composition on the true frame center under captions. Quotes the rail+embed model from embedded-captions and constraint

Write tests that pin real behavior instead of implementation details, config values, or lucky samples. Use when adding tests for new behavior, writing a regression test, fixing a brittle or flaky test, reviewing a test diff, or when a test breaks after a refactor or config change that didn't change behavior.

Break large features into independently verifiable, human-reviewable slices under a feature directory in specs/. Use for risky or multi-step feature work that needs upfront questioning, API seams, browser-playable checkpoints, HTML visualizations, screenshot gates, staged implementation plans, recursive fog-of-war reslicing, or proactive research into reference implementations/best practices before slicing. Pairs with your project's verification harness and screenshot gates (the browser checkpoints), [refactor-clean](../refactor-clean/SKILL.md) (review the materialized spec so the plan describes one-owner architecture, not the feature bolted on), [screenshot-critique](../screenshot-critique/SKILL.md) and [compare-screenshots](../compare-screenshots/SKILL.md) (the visual gates), and a code-review pass (audit each slice before it lands).

Create or revise agent skills. Use when adding a new skill file, renaming a skill, simplifying an existing skill, improving trigger descriptions, or deciding what belongs in a skill versus references, scripts, assets, or ordinary docs.

Write and edit project docs (README/markdown) as a glossary of principles, not a mirror of the code. Use when creating or revising a README; when a doc enumerates exact scenes, scenarios, helpers, class ids, file lists, or command/flag matrices the code already holds; when trimming narrative or changelog out of a doc; or when deduplicating overlapping docs and wiring a root doc to its sub-docs.

Build AI-powered software with TypeSafe: small units of AI intelligence you can use like programming primitives. Its System One models, including Jev, turn natural language and application state into typed judgments and probabilities that code can combine. Use when a feature needs programmable common sense, when brainstorming what AI could make possible in an app, or when an LLM prompt-and-parse step could become a structured decision. Applications include routing, ranking, extraction, verification, and interactive experiences; these are starting points, not the limits. Read live docs and cookbooks to find useful patterns and discover new combinations.

Use the unprimed sub agent as a second set of eyes before accepting visual work — MANDATORY before declaring any user-reported visual bug fixed or claiming a visual change verified; primed eyes pass defects fresh eyes catch.

Closeout pass on a finished substantive change — sequence refactor-clean (shape), code-review (diff), and write-docs (docs) so all three run, in the order where each feeds the next. Use after implementing a slice or feature and before calling it done or committing; when the user asks for a final review, a cleanup-and-document pass, or to "review everything." For a diff-only audit, reach code-review directly.

Refactor cleanly instead of layering sediment. Use for rename requests, stale or misleading names, changed scope, or when a change reveals duplicated concepts, local adapters, obsolete owners, redundant inputs or guards, compatibility wrappers, parallel abstractions, an over-large module that has accreted many responsibilities, or "just tack this on" pressure in any code area.

Open image shots (screenshots, snapshot baselines, montages, renders) in ONE macOS Preview window so the user can eyeball them. Use when the user asks to open / show / pull up / review shots in Preview or Finder, or right after producing visual artifacts the user should look at before signing off.

Best-practice rulebook for a marketing site's pages, organized by page class — campaign landers, product pages, programmatic SEO trees, editorial. Use when writing or updating a landing/marketing page, wiring a campaign lander to an attribution identifier, changing site-wide nav/footer links, touching sitemap/robots/agent-readable surfaces, or auditing a public site's indexing, reachability, and conversion setup.

Make a showreel-grade motion-graphics launch video, with music, that introduces a CLI, library, or product from its repo. Use when asked for a launch, promo, sizzle, or showreel video, an animated intro for a project, or a video to post on X/social.

Use for questions about how, why, or where behavior works in a repository, including questions that name a function or setting. Start behavioral discovery with jg before broad text searches or git history. When delegating repository discovery, instruct the subagent to start with jg. Jevgrep returns relevant files and source excerpts. For exact symbol definitions, string matches, or filenames, use grep or file search instead.

Run [implement-spec](../implement-spec/SKILL.md) with Codex doing the implementation passes while you orchestrate, integrate, and review. Use when the user asks to implement a spec with Codex, or to run implement-spec with Codex as the implementer.

Implement an existing spec through committed passes. Use for long or multi-pass specs that need maintenance checkpoints to periodically clean code, handoffs, priorities, and plan bloat before drift accumulates.

Hand a spec that is mid-implementation to another machine or another agent.

Guide the user through a quadrant walk that maps the unknowns of a task — open by listing the known knowns, then work through known unknowns, unknown knowns, and unknown unknowns one stage at a time, ending with a complete four-quadrant map in the user's hands. Use when a request is ambiguous or underspecified, the codebase or domain is unfamiliar, the user will "know it when they see it", a reference implementation must be understood before porting, mid-build deviations from the plan need capturing, or a finished change needs buy-in or verified understanding before merge. Pairs with [write-spec](../write-spec/SKILL.md) — walk the quadrants to burn off fog before slicing, and feed the finished map into the spec.

Eval and improve a skill against golden cases — run the target skill blind in a fresh, context-free subagent on each example input, grade the artifact against the expected outcome, and let the gaps drive the edits. Use when the user wants to test/eval/improve/harden a skill, says "this skill keeps producing X / keeps missing Y", or hands a skill plus example input→expected-output pairs. Pairs with [write-skills](../write-skills/SKILL.md) (the authoring principles every fix obeys).

Explain a technical spec or proposed change in plain language: what problem it solves, how the solution works, and every schema change.

Explore UI and visual design with image generation, then iterate the real implementation against the user-selected concept. Use when brainstorming visual directions, offering mockup variants, or implementing an approved generated design; do not stop at the mockups or the first code approximation.

Use the local Codex CLI as an independent second agent. Two branches — (1) proactively run `codex review` for a second opinion after completing a substantive change, before presenting it as done or committing; (2) delegate a well-defined implementation task via `codex exec`, ONLY when the user explicitly asks for Codex to do it. Also covers how to prompt Codex.

Use Claude Code as an independent `claude -p` subagent when the user explicitly asks for Claude, wants a second-agent opinion from Claude, or asks to delegate a well-scoped task to Claude. Supports selecting `--model` and thinking/effort level with defaults of `opus` and `high`.

Run fast, progressive experiments to map tunable parameters and their effects. Use when optimizing code, prompts, configurations, or other artifacts against a goal or benchmark, or when a long research run needs a planned hypothesis queue, focused trials, and durable findings.

Audit whether tests earn their maintenance cost and which suite owns each contract. Use when pruning redundant tests, investigating implementation coupling or test-only production hooks, reviewing the value of proposed coverage, or auditing an entire subsystem. Use write-tests to implement the resulting test changes.

Audit and prioritize performance work through bounded-work and forward-progress checks. Use when asked to find performance issues, investigate freezes/500s/OOMs, review retries or polling for no-progress loops, rank a performance backlog, or implement the next simple performance fixes.

Audit the choices an implementing agent made, not its diff — a pure decision audit that traces the session's history into a choices ledger, changes no code, and never blocks an unsupervised run. Working code still embeds architecture the user never chose; surface it because future work inherits it. Use when the user wants to review the decisions the AI made on their behalf, before merging or committing AI-implemented work, when integrating a delegated subagent's pass, or when a fix "works" but might be a point fix.

Audit or rewrite AGENTS.md so it holds only lasting principles. Run it only when the user asks for it by name; never invoke it on your own.

Clicks through every flow of an app clone and tests it for bugs: a test plan generated from the recon flows with happy paths and edge cases, Playwright end-to-end tests where possible, a browser click-through where not, and bug reports in a fixed format with severity, steps and evidence. Use when the user says "test my clone", "find bugs", "QA this", "click through everything", "write e2e tests", "does it work", or after /replica-build or /replica-backend.

Reverse-engineers any app into a recon map: screen inventory, user flows, component list, inferred data model and a feature matrix, from public pages, screenshots, app store listings, help docs and the user's own account. The first step of the Replica pack. Use when the user says "clone this app", "reverse engineer X", "how does X work", "map out X", "what screens does X have", "I want to build my own version of X", "copy this app", or pastes an app's URL or App Store link and wants to rebuild it.

Writes the launch for an app clone: the landing page built on the positioning angle, pricing set against the original's public pricing and its users' complaints, and the App Store and Google Play listing linted against the stores' limits and copycat rules. Use when the user says "landing page", "pricing", "how much should I charge", "App Store listing", "store screenshots", "launch plan", "Product Hunt", or after /replica-brand.

Researches the app being cloned and reads what its real users say in public reviews (App Store, Google Play, G2, Capterra, Trustpilot, Reddit, Hacker News, its own feature-request board), then ranks what they hate, what is missing and what is unsolved, turns it into a fix plan for the clone and a positioning angle you can sell. Every quote is real, verbatim and linked. Use when the user says "what do people hate about X", "read the reviews", "how do I make mine better", "find the gap", "what features are missing", "how do I position this", "make it sellable", or after /replica-diff.

Compares an app clone against the original: a feature parity score from the feature matrix (weighted by must, should, could) with the missing list in build order, plus a screenshot layout diff that ignores colour so a rebrand does not count against you. Two standard-library Python tools. Use when the user says "how close is my clone", "compare it to the original", "what's missing", "parity check", "diff the screens", "is it ready", or after /replica-test.

Rebuilds an app's design system for a clone: colour roles, type scale, spacing, radius, shadows and every component with its states, as design tokens plus component specs, with original assets instead of the target's logos, icons, illustrations or licensed fonts. Includes a WCAG contrast checker. Use when the user says "match the design", "rebuild the design system", "get the colours and fonts", "make it look like X", "design tokens for my clone", or after /replica-architect.

Ships an app clone live on the user's own domain: a preflight gate (tests green, parity must-haves done, rebrand sweep clean, listing linted, legal pages up), production database and env vars, the host, DNS records for the domain and for email, Stripe live mode, OAuth redirects, monitoring, and mobile builds to TestFlight and Play. Use when the user says "deploy it", "ship it", "put it live", "connect my domain", "go to production", "publish the app", or after /replica-launch.

Rebuilds an app screen by screen from the recon map: app shell first, then the core flow as a vertical slice, then every screen with all its states, ticking off the feature matrix as it goes. Writes every line fresh, never the original's code, assets or copy. Use when the user says "build the clone", "build screen S07", "start building", "rebuild this screen", "implement the booking page", or after /replica-design.

Names and rebrands an app clone so it is the user's own: name candidates with the trademark, domain, store and handle checks to run, a new palette checked for contrast, a logo brief, a voice guide, and a sweep tool that finds anything of the original left in the codebase. Use when the user says "name my app", "rebrand the clone", "make it mine", "pick a name and colours", "logo brief", "brand voice", "check for leftovers", or after /replica-entrepreneur. Always runs before /replica-launch.

Builds the backend of an app clone: auth, database migrations and access rules, payments with Stripe, email, background jobs and third-party integrations through official APIs only, plus a security checklist. Use when the user says "add login", "set up auth", "wire up the database", "add payments", "connect Stripe", "add Google Calendar", "send emails", "backend for my clone", or when /replica-build is running on fake data.

Plans the stack, database schema and API for an app clone, from the recon map replica-recon wrote. Picks boring, managed tech, turns the inferred data model into real SQL with indexes and access rules, lists every route by flow, and orders the build as a thin vertical slice first. Use when the user says "plan the clone", "what stack should I use", "design the database", "write the schema", "plan the API", "architecture for my version of X", or after /replica-recon finishes.

Check on, troubleshoot, or uninstall an installed Insomnia (the macOS menu bar app that keeps a Mac awake for a timed session). Use when the user asks why Insomnia shows a warning, whether a session cleaned up after itself, why the hotspot password stopped working, or how to remove it.

Install or upgrade Insomnia, the macOS menu bar app that keeps a Mac awake for a timed session, from a GitHub release or from source. Use when the user asks to install, update or reinstall Insomnia.

论文精读：把一篇论文讲透，不只是译出来。用户说“讲透这篇”“精读这篇”“这篇到底是怎么想出来的”“x-ray 一下”时使用——先通读全文含附录，还原作者真实的思考起点、押的那个赌注和证据落在哪，把每个符号和关键公式落到能自己动手算的小例子上，再怀疑式地过一遍超参、消融、基线和数字。产出写进 EasyRead 的页边讨论条目并锚到对应段落，也可以另存一份长文 Markdown。用于 精读论文、讲透一篇论文、还原作者思路、这篇论文哪里有问题、paper-xray、$paper-xray。不是摘要、不是逐段翻译、不是科普文。

论文共读：用 EasyRead（E:\CursorProject\easyread）这个本地工具读论文。用户给一篇 PDF 或 arXiv 编号时，导入文献库、翻译成中文（后台引擎或对话里的 agent 亲自译）；之后边读边讨论，agent 读用户在页面上的笔记和提问，把回答、解释、原文核对提示追加到对应段落旁。用于 论文共读、读论文、翻译论文、导入论文、回答我在论文里提的问题、paper-reading、$paper-reading。不是摘要、科普文或 explainer 页面。

Explore ontology questions, grounded concepts, evidence, actual task results and evolution outcomes in a read-only interactive version explorer.

Improve an existing ontology using real project tasks and grounded exploration, run missing baselines, validate candidates and automatically show the outcome. This skill is triggered when the user requests to "self-evolve the current ontology layer system". Autonomously evolve an ontology-layer system through a four-step loop of diagnosis, attribution, patching, and Parent/Candidate evaluation until a reproducibly better version is obtained, or external conditions prevent trustworthy continuation.

Build an initial ontology from data and analytical goals, prepare relevant questions from supplied needs and project history, and automatically show the evidence-grounded result.

Point at something on the human's macOS screen with a big arrow and a sign ("Franz, click HERE") using the bigarrow CLI. Use whenever the human has to look at, click, type into or choose a specific spot, button, field, window, tab or menu: show me where, point to it, which button, where do I click, zeig mir wo, guide me through this UI, the human must approve or sign something, or you need the human's hand and they may not be looking at the terminal. Mandatory when it is time-critical: a code that expires, a login or payment page that times out, a build waiting on the human.

Compile a repository's instruction files (AGENTS.md, CLAUDE.md and friends) into an Abide rubric, then validate and calibrate it.