REA
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
This repo is a collection of 111 skills
A local-first desktop workspace for persistent AI teammates. Run Codex, Claude, and Grok with dedicated workspaces, task queues, file sharing, browser control, and agent-to-agent collaboration.如果你是作者,可把这段代码贴到 README。徽章会链接回本页。
[](https://mcp.popnx.com/en/skills/nightly-labs-openbot)在同一工作中把 AI 连到服务的 MCP
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
Codebase intelligence for TypeScript and JavaScript, exposed as structured MCP analysis tools.
MCP server for terminal commands, file operations, and process management
同类工作,排序与 Skill 列表相同
| 本项 openbot | rea | andrej-karpathy-skills | graphify | MoneyPrinterTurbo | |
|---|---|---|---|---|---|
| 作者 | nightly-labs | morluto | multica-ai | Graphify-Labs | harry0703 |
| Skill 数 | 111 | 1 | 1 | 1 | 1 |
| GitHub 星标 | 8,782 | 467,235 | 234,125 | 149,873 | 144,332 |
| 最近代码更新 | 2026年11月18日 | 2026年11月18日 | 2026年4月20日 | 2026年11月18日 | 2026年11月18日 |
| 许可证 | – | MIT | – | Apache-2.0 | MIT |
nightly-labs 提供的 111 个 Skill(指令集),AI 读取后照做 · 不是 MCP server
Install only the ones you use. Open a skill to see how to install it
This set has 111 skills, too many to install at once: every installed skill’s description loads into every chat, so the AI picks the wrong one more easily. Pick only the ones you use from the list below
Test a live web app against an acceptance checklist in a real browser and return PASS, FAIL or BLOCKED with screenshots and repro steps. Use before a release, after a deploy, or for a scheduled smoke check of a saved site.
Audit a page in the browser for keyboard path, focus, names and roles, contrast, headings, forms and errors, images and motion. Each finding has a severity (blocker, serious, minor), a WCAG 2.2 success criterion number and evidence. Use before a release or when a user reports a barrier.
Score a company against the user's ideal customer profile with a fixed rubric, using evidence from public company pages. Lists disqualifiers, gaps and discovery questions. Named people come only from public business pages, with no private details. Use before outreach or a first sales call.
Audit the recent chats of local OpenBot agents for user friction, propose new skills, agents or routines from repeated corrections, or run the standing transcript-healthcheck routine.
Design and create a new OpenBot agent. Ask a few preference questions, write a tight profile, call create_agent, then verify it with read_agent. Coding agents use pstack poteto-mode as the quality bar. On a fresh install, run Eggbot Getting Started first.
Write short, copyable alt text for images: the purpose of the image in its context first, 125 characters or less by default. Handles decorative images, charts, images of text, logos and linked images. Returns one copy block per image.
Turn an API or a repeated workflow into a tested local OpenBot skill. Read the docs, sample one real response first, keep keys out of files, write the smallest SKILL.md and a script only when needed, test it, then register it. Use when the user repeats the same request against a service.
Sketch types, signatures, and module structure before code, then stay in the loop while implementation fills in. Use for /architect, 'architect this', 'design this', or non-trivial work where jumping to code would lock in the wrong shape.
Spawn N parallel candidates at the same task, pick a base, graft the strongest parts of the losers into it. Use for /arena, 'arena this', 'throw it in the arena', or when one attempt at a non-trivial artifact would lock in the wrong shape.
Use for "automate me", "create/update/refresh my -mode skill", "turn/capture my preferences or working style into a skill", or wanting agents to follow how the user works. Drafts or revises a personal -mode skill via create-skill + unslop, optionally pulling fresh evidence from recent transcripts.
Babysit a pull request through the NorbiAI review loop until no valid issue remains. Fix valid findings, rebut false positives with checkable evidence, and re-request review. Use when asked to babysit a PR, handle NorbiAI comments, or clear the NorbiAI review gate.
Vet a perf measurement (limiter, tuning, limits, errors, repeatability, relevance, and whether the work happened) before you report or act on it. Use when you run a benchmark or report a speedup or regression you measured.
Install and configure vendored Biome anti-slop GritQL rules in a local TypeScript or JavaScript repository. Use when a user asks to add, configure, update, or migrate Biome anti-slop lint rules.
Find what a change could break somewhere else before it ships, beyond the diff, and prove the one fact it's safe because of by running real code instead of writing it up. Use for 'blast radius of X', 'what could this break', or reviewing a small diff you don't trust.
Restate the last message in plain human language, with no jargon.
Turn a bug report, logs or a failing test into a reproduction, ranked causes, one diagnostic step that separates them, the smallest fix, and a verified result. Use when something fails and the cause is not yet known.
Turn a design frame into a spec a developer can build from: components, layout and spacing, tokens and where values drift from the design system, states, motion, content rules and open questions. Works from an installed Figma plugin, a signed-in Figma tab, or a screenshot.
Score a sales call transcript from 1 to 5 on seven named criteria, each with a timestamped quote as evidence, then give the top 3 fixes for the next call with a better line to try. Works from a pasted or uploaded transcript, or a transcript in a signed-in call recorder tab. Never invents quotes.
Explain what drove a change in campaign results and propose 1 to 3 measurable experiments. Reads an installed analytics plugin such as PostHog, a signed-in analytics tab, or an export. Checks units, dates and attribution first, and keeps observations apart from causal hypotheses. Never changes camp…
Keep a private care log for plants, pets, filters and appliances: what needs doing, how often, when it was last done and when it is next due. Add items by chat or photo, mark tasks done, list what is due, and show a journal on request. The log stays on the user's computer.
Review a code change for bugs, regressions and missing checks. Each finding gets a location, a trigger, the consequence and the smallest fix, ranked by severity. Use for a pull request, a branch, uncommitted work or a pasted diff.
Compare products or vendors against criteria chosen from the decision. Every table cell has a source, prices carry the date they were checked, and a fact that was not found is marked unknown. Use to choose a tool, to position a product, or to check a claim about a competitor.
Find the mistakes agents keep repeating in this repo and make each one impossible. Try architecture first, then types, then a lint whose error names the fix, then a test, and write docs last. Prove each check fails on a real past mistake. Repeat this each time the operator corrects you. Use for /co…
Generate a project-local verification skill that drives your app the way a user does — any language, framework, or platform. Use for /create-verification-skill, "make a control skill for this repo", or when a project has no scripted way to prove UI/CLI/service behavior.
Set up a Leaseweb VPS as the OpenBot server of one customer (an enterprise or custom server), with the customer as the owner. Use when asked to set up, assign or give a VPS or a dedicated server to a customer or an email, to find a Leaseweb VPS that was just bought, or to continue a custom server s…
Write a one-screen morning brief from the signed-in calendar and inbox and installed task tools: today's meetings with prep notes, commitments due, replies waiting, and conflicts. Drafts replies and never sends. Use each morning or before a busy day.
Check a dataset before analysis: row counts, duplicates, missing values, units and currency, date ranges and time zones, joins, and outliers. Returns a pass, warn or fail table and names which analysis each problem blocks. Use before a report, a model, or a decision that rests on the data.
Add a brief, concrete self-check to consequential recommendations, estimates, plans, or interpretations when unverified assumptions could change what the user does.
Give new or substantially redesigned interfaces a deliberate visual direction grounded in their subject, audience, content, and purpose. Do not use for minor styling fixes.
Run first-conversation onboarding when you are a newly installed copy of Dr. Eggbot, the agent designer. Set up pstack once, make sure the two healthcheck routines exist, and save the owner's preferences.
Answer a research question with sourced, dated claims. Each claim has a source title, publisher, publication date and access date, and a label: Verified, Reported or Inference. Conflicts between sources stay visible. Use for fact checks, market questions, and any answer that a decision or a publish…
Design an auditable playbook when no narrower one fits: a large migration, an ambitious multi-part change, or work a human reviews after stepping away. Scales rigor to the task, runs a hypothesis loop, and logs decisions. Use for /figure-it-out or 'figure it out'.
Export the user's Grok Bot agents into one .zip file that OpenBot imports, with instructions, avatars, skills, routines, memories, group chats, and optional workspace files. Use when the user asks to export, move, or migrate their agents to OpenBot.
Export the user's Hermes Agent profiles into one .zip file that OpenBot imports, with each profile's persona, memories, skills, and cron jobs. Use when the user asks to export, move, or migrate their Hermes agents to OpenBot.
Turn role requirements into competencies, structured questions for each interviewer, a 1 to 4 scale with behavioral anchors, and an evidence-only debrief. Use when you plan an interview loop or summarize interview feedback.
Use for "how does X work", code walkthroughs before changing something, and placement / ownership / layering questions ("where should this live", "which package owns this", "is this the right layer"). Explains subsystem architecture, runtime flow, onboarding mental models. Use why for motivation.
Review a screen and return ranked, concrete fixes across hierarchy, layout, type, color and contrast, copy, states and accessibility. Works from a screenshot, a page in the browser, or a Figma frame. Never edits files.
Draft status reports, leadership updates, project announcements, FAQs, newsletters, and incident communications for an internal audience.
Use for "interrogate", "adversarial review", "multi-model review", "challenge this", "stress test this code", "find blind spots", or "tear this apart". Multiple LLM reviewers challenge changes from independent angles.
Catch the user up on everything in this chat since their last message, in plain words. Give each action item an executive summary with enough context to make the call. Use for /ketchup.
Write release notes from merged work, with a launch checklist that has owners and acceptance checks. Keeps done, planned and blocked work apart. Use before a release or a launch announcement. It never publishes.
Periodic pass that keeps a project's verification skill and feature map honest: parallel source readers per feature, one live session driving every feature, at most one PR of proven corrections. Use for /maintain-verification-skill or "audit the verify skill".
Use when building a custom UI (page, dashboard, buttons) that should wake a Grok Bot over a webhook, when the user must provide a webhook sender key, or when exposing that UI on Tailscale.
Turn meeting notes or a transcript into decisions, commitments with owner, due date and quoted evidence, open questions and conflicts. Use after a meeting, before a follow-up, or when someone asks who agreed to what.
Spawn Comment Sicko, fix accepted findings, and offer encodings for claimed constraints.
Keep structured information in the shared database that every OpenBot agent reads and writes. Use when you must look up, update, or count records such as people, tasks, messages, or results across turns.
Build, publish, replace, list, and delete small static websites hosted by OpenBot. Use when the user asks to host or manage a site on openbot.site.
Create or revise a reusable local OpenBot skill when the user asks to save a workflow as a skill. Use the local skill tools to register it. Do not use for ordinary one-off tasks.
Query OpenBot product analytics (events, sessions, profiles, funnels, retention, active users, Search Console, and SQL percentiles such as p50/p90/p99) from the self-hosted OpenPanel server, and make or change its dashboards and reports. Use when debugging with real usage data, checking whether an…
Draft a first business message built on one verified reason to write and one clear request. Length fits the channel. Without evidence, it writes an honest general draft instead of fake personalization, and it lists the claims the user must confirm. It never sends.
Guides users through pstack setup, /poteto-mode, and picking the skill, playbook, or principle for a task. Type /poteto-help with a question.
poteto's agent style for concise, detailed responses, deliberate subagents, unslopped prose, simple code, and verified work. Use for poteto, /poteto-mode, or requests to work in this style.
Track the open pull requests of a named repository, classify each one, fix failing CI and review threads on the PR branch when the user allowed it, and ask the user only to merge the ones that are ready. Use for a one-time pass or a scheduled check.
Apply when two or more fixes that share one premise have failed the same gate. Take a census of which actors hold the imbalance before the next fix, then question the premise instead of writing another fix that assumes it.
Apply when wiring validation, error handling, or framework adapters. Concentrate guards at system boundaries (CLI, config, network, external APIs); trust internal types and keep business logic in pure functions.
Apply to any non-trivial work, not just bulk work: edits, migrations, analyses, checks. Build the tool that does it or proves it (codemod, script, generator, or a skill your subagents follow) instead of working by hand. The tool is the artifact a reviewer can rerun.
Apply when you catch yourself writing the same instruction a second time, or notice a recurring correction. Encode the rule as a lint, metadata flag, runtime check, or script instead of more text.
Apply when facing a novel UI interaction or architectural decision with no precedent in the codebase. Build 2-3 competing prototypes and compare side by side before committing.
Apply when product, UX, or feature-scope tradeoffs come up. Choose user delight over implementation convenience; ship fewer polished features over more rough ones.
Apply before you trust, report, or act on a number you measured: a speedup, a regression, a throughput, a latency, or an eval result. Find what limits it, and rule out that it measured something other than the work you think.
Apply when debugging. Trace each symptom to its root cause and fix it there; reproduce first, ask why until you reach it, resist nil-check guards that silence crashes.
Apply before writing logic: choosing core types and data structures, sequencing scaffold-vs-feature work, asking what concurrent actors share. Get the data structures right so downstream code becomes obvious.
Apply when context is filling up: large outputs, long files, repeated reads, fan-out planning. Route bulk to subagents; keep summaries in the main thread, not raw payloads.
Apply when refactoring, evaluating diff size, or tempted to add abstractions, layers, or signal threading. Bias toward deletion and the smallest change that solves the problem.
Apply when designing commands, lifecycle steps, or processing loops that run amid crashes, restarts, and retries. Converge to the same end state regardless of partial prior runs.
Apply when introducing a new internal API while old callers still exist. Migrate callers and delete the old API in the same wave instead of preserving compatibility layers.
Apply when reviewing or shaping code that's hard to trace. Count layers between question and answer, and hidden state in the reader's head; collapse one-caller wrappers and shrink mutable scope.
Apply when writing stateful logic, or when code branches a lot or repeats a shape assumption across files.
Apply when tempted to ask 'should I do X?' on reversible work.
Apply during planned rewrites and migrations with explicit phase boundaries. Converge on the target architecture; don't preserve smooth intermediate states with throwaway compatibility code.
Apply after completing a task, before declaring done. Verify against the real artifact (run the feature, read the actual value, inspect the diff), not a proxy, self-report, or 'it compiles.'
Apply when integrating a new requirement into an existing design.
Apply when concurrent actors might write to the same file, branch, key, or state object. Eliminate the sharing first; serialize structurally only when one shared writer is a real invariant.
Apply to multi-step work (sweeps, migrations, runs of similar edits) and to how you stack commits and PRs. Break work into small units that each end in a verifiable state, check each before the next, and order delivery so the sequence proves itself to a reviewer.
Apply when sequencing an addition, refactor, or rewrite. Remove dead code, redundant validators, and stub references first, then build on the simpler base.
Apply when you write, change, or keep a test. Call the code as its users do and assert the exact result they see. Rewrite or delete a test that would pass if every imported function returned undefined.
Apply when designing types, reviewing a function signature, or writing code in any statically-typed language. Make illegal states unrepresentable, brand semantic primitives, parse external data at boundaries, refuse to lie to the compiler, exhaust variants, derive from authoritative schemas.
Write a dated scope document: problem, what is in, what is out, how completion is measured, the first deliverable, risks, decisions needed and owners. It stays a draft until the user approves it, and no build work starts before that.
Build several different design directions (3 by default) for a feature or issue as Storybook stories, show them to the user, and let the user choose one. Use when the user writes `$prototype `, "prototype X", or asks for design variants, options, or directions before implementation.
Turn an idea into the smallest clickable static prototype, click through it in the browser, and return screenshots of the click path and a short video. Publishes to a public site only after the user says publish. Can open a draft pull request instead when the user asks.
React-to-SolidJS transformation patterns for porting shadcn-style React components, examples, docs snippets, and blocks into Zaidan. Use when converting React TSX, React hooks, Radix/Base UI/shadcn patterns, Next-style code, or React-specific dependencies to idiomatic SolidJS.
Reconstruct your recent working context from your own chat history, live state, and the shared record (user reports, prior fixes, incidents), then hand back a tight current-state brief. Use for 'recall my work on X', 'catch me up', 'what have I been working on', 'where did I leave off', before star…
Record a short video of a change in the real local app and attach it to the pull request. Use after a UI change, a new flow, or motion, when the user writes `$record-demo`, or asks for a demo, a video, or a recording for a PR.
Find recurring charges, price changes, trials that convert, duplicate charges and statement credits that expire, from pasted or downloaded card and bank statements. Shows the statement lines as evidence and lists renewals due soon. Never cancels or contacts merchants.
Spawn three parallel review subagents over the active transcript, surface learnings, and route each to a concrete edit on an existing skill. Use when the user says reflect.
Audit a pending OpenBot release for upgrade and data-loss hazards before the version is bumped or tagged. Use when asked to cut a release, create or bump a version, prepare a tag, or check whether a change is safe to ship to installed users.
Produce a repository-grounded application security threat model with assets, boundaries, abuse paths, priorities, and mitigations. Use only when threat modeling is explicitly requested.
Score inbound event, sponsorship, speaking and partnership requests 0 to 2 on criteria the user sets once, turn the total into yes, maybe or no, and draft a reply for each. Keeps a log and never sends. Use when requests pile up in a channel or inbox.
Audit the routines of all local OpenBot agents for token waste, offer a fleet checkup after install, decide routine cadence, or run the standing routine-healthcheck routine.
Turn a keyword list or Search Console data into keyword groups by intent, chosen targets, and writer-ready briefs for search results and AI answers. Each brief has the question, angle, outline, sourced facts, internal links and an answer-first summary. No invented search volumes.
Review or write Python, JavaScript, TypeScript, or Go code using secure defaults. Use only for explicit security guidance, security reviews, or secure implementation requests, not ordinary code review.
Configure which models pstack uses per role and at what reasoning budget. Detects your available models and writes an always-applied rule that overrides the skill defaults. Use for /setup-pstack, "configure pstack models", "pstack budget", or changing pstack's model choices.
Ship the current branch to main. Review the diff, fix findings, open the pull request, babysit the NorbiAI review and CI until green, then squash-merge. Use when a feature is done and the user asks to ship it, open and merge a PR, or get the change to main.
Keep a reviewable decision trail for long-running or unattended work: a TSV log with one row per decision (what, why, evidence, result). Local by default; commit it when a reviewer needs the trail to trust the result. Use for /show-me-your-work, autonomous or multi-phase runs, or work a human revie…
Check and deploy Signal (the `remote-api` service in `remote/`) to production from a release tag. Use when asked to deploy, redeploy, update or roll back Signal, to check that a Signal deployment has all its secrets and no regressions, or when a release tag changes `remote/` or `packages/contracts/…
Start one or more local OpenBot dev instances and smoke test the features that the current changes or a pull request add. Use when asked to smoke test, QA, or check a change or PR in the real app, including two-client team flows, the browser client, and the iOS simulator.
Smoke test the current branch, then ship it to main only when all smoke items pass. Use when the user asks to smoke and ship, or to test a change in the real app and then merge it.
Build an inventory of software spend from bills, invoices, card exports or billing pages, then find unused seats, duplicates, overlapping tools, cheaper plans and near renewals, each with evidence and the saving arithmetic. Drafts vendor messages for review. Never cancels, pays or sends.
Draft a customer reply that states what is known and the next step, plus a separate internal note and, when needed, an escalation package with repro steps, account facts and impact. No refunds, exceptions or dates without a policy that allows them. Sends only on an explicit request.
Fan out N parallel workers, drain them, and return one report. Use for /swarm, 'swarm this', or parallel coverage, races, gauntlets, and exploration.
Use only when the user explicitly asks for TDD, a failing test, or a regression test, OR when the bug has an obvious cheap local test target. Skip when the test path is unclear, expensive, integration-heavy, or not requested.
Explain a body of work plainly so a person actually understands it. Runs the `how` and `why` skills and weaves what they find into one clear explanation. Use for 'teach me this', 'help me really understand X', 'explain this change or subsystem to me'.
Layered technical-writing standard: Diátaxis structure, Google developer style sentences, STE instruction rules, Global English syntax. Use for /technical-writing or when writing or reviewing docs, RFCs, readmes, PR descriptions, or commit messages.
Triage one GitHub issue, bug report, or feature request. For a bug, decide if it is valid, rate its UX impact on casual users, find the root cause, and fix it. For a feature request, decide if it is necessary, bloated, or over-specific, and score its value, UX, and priority. Push back with evidence…
TypeScript best practices. Use when reading or editing any .ts or .tsx file.
Cut AI tells from any writing.
Edit a draft so that it sounds human and sounds like the user. Learns the voice from 1 to 3 samples of the user's own writing, removes filler and stock phrases, keeps every fact, number and claim, and adds no new facts. Returns the edited text and a change log with the reason for each change.
Use for 'why does X work this way', 'why we picked Y', design rationale, regressions, postmortems, or data-backed thresholds. Discovers available MCPs and queries each evidence category (source control, issue tracker, long-form docs, real-time chat, infrastructure observability, error tracking, pro…
Track a recurring operations list (shipments, facilities issues, key dates) in a shared table, keep a short playbook for each item type, and write a weekly digest of overdue, due, done and new items. Use for office or team operations that repeat every week.
Use when you reach out to or talk with another of the user's OpenBot agents: deciding whether to involve one, writing the handoff, reading its reply, answering a teammate, group chats, and stopping. Also a general guide to a warm, concise, careful assistant style.
Install and configure the Zaidan component registry in an existing SolidJS codebase, and recommend Zaidan components for a product or interaction need. Use when setting up Zaidan, adding Zaidan components, fixing consumer-side registry configuration, or deciding which Zaidan components to compose f…
MCP = 手
把 AI 连到外部服务或数据,例如读取 Notion 或开发票。需要有程序或服务器在运行。
Skill = 手册
教 AI 怎么做某项工作的文本文件,不连接任何服务。
有些会重叠:少数 Skill 带有调用工具的脚本,少数 MCP 只提供知识。请从你想完成的工作出发。
Claude CodeClaude (เว็บ/แอป)CodexCursorGitHub CopilotOpenCode
依据开发者 README 和 repo 中的文件夹
This repo has 111 skills. Pick one from the list above; its page shows every way to install it, including letting the AI install it and a ZIP
提问、分享技巧或报告无法使用。按“有帮助”数量排序。