agentsclimarketplace

The catalog

608,666 rows, by stacks then stars

  • Codex project memory

    Retourflow/project-memory-for-codex/skill/codex-project-memory Skill

    1★ repo

    Maintain a minimal, token-aware, verified project memory for Codex repositories. Use when initializing project context files, resuming work across sessions, changing architecture or project goals, recording decisions, updating current state after implementation, or reducing bloated and stale agent documentation.

  • Cloudflare drop

    ofoxai/skills/skills/cloudflare-drop Skill

    1★ repo

    High-standard open-source agent skills from OFOX AI — one monorepo. Install: npx skills add ofoxai/skills

  • Index365 add project

    index365usa/agent-skills/skills/index365-add-project Skill

    1★ repo

    Request AI-Readiness and Marketing Signal scans and use prioritized findings from your coding agent.

  • Anti slop engineering

    Anrahya/skills/anti-slop-engineering Skill

    1★ repo

    Senior engineering guardrails for writing, reviewing, debugging, refactoring, or planning code. Use when an AI coding agent is working in a repository or advising on implementation choices and should stay simple, surgical, evidence-driven, product-aware, and willing to challenge weak technical suggestions instead of blindly following them.

  • Lys teach

    linkyourskill-ai/lys-skills/lys-teach Skill

    1★ repo

    Teach the user a new skill or concept, within this workspace.

  • Architect

    vasu-devs/Forge/skills/architect Skill

    1★ repo

    Design the shape of a non-trivial module, API, or system before implementing it. Use when a design decision has more than one reasonable answer, when introducing a new abstraction or interface, or when getting the shape wrong would be expensive to undo.

  • Figma to email html

    AzalDaniel/FigForge/skills/figma-to-email-html Skill

    1★ repo

    Converts a Figma design (frame, component, or exported SVG) into faithful, email-safe HTML/CSS without paid tools like Anima, Kombai, or Postcards. Use when someone shares a Figma export or URL and wants pixel-accurate HTML for an email, newsletter, or template; wants to bake gradients, shadows, or glows into images; extract icons/photos as transparent PNGs; or make markup render the same across Gmail, Outlook, and Apple Mail. Drives a forensics -> asset-extraction -> effect-baking -> render+diff -> robustness-lint loop using the open-source `figforge` Python package as the engine.

  • Cloudflare domain registration

    nachoal/my-skills/plugins/claude/my-skills/skills/cloudflare-domain-registration Skill

    1★ repo

    Check availability and register domains through Cloudflare Registrar. Use when asked to buy, purchase, register, check, price, or verify a domain on Cloudflare Registrar with CLOUDFLARE_REGISTRAR_TOKEN.

  • Me os skills

    taohaowei/me-os-skills Skill

    1★

    一个本地优先的私人助手框架——亦师亦友,一个统一入口处理工作、心理、家庭、健康、个人成长所有维度的对话。 触发场景:用户讨论团队管理 / 项目方向 / 向上汇报 / 1on1;情绪困扰 / 家庭矛盾 / 工作压力 / 亲密关系 / 焦虑 / 育儿;想倾诉 / 寻求心理支持;记录就医/用药/体检/症状;个人成长 / 决策困境 / 人生迷茫;自我认识;说"先聊聊"或没有明确话题;以及"老师" "教练" "咨询" 等称呼。 框架(镜头 / 协议 / 方法论)在本技能目录里;用户的真实档案数据存在 $ME_HOME(默认 ~/.me)。Invoke via /me.

  • Lookdev studio skill

    connerkward/lookdev-studio-skill Skill

    1★

    When the user says "lookdev" (or asks to tune/dial/iterate on the look of something, OR to review/edit/annotate a blog post, doc, copy, or media), stand up a local interactive studio — live controls (sliders, pickers, drag handles) for visual params, OR direct inline editing + highlight + margin comments + media annotation for prose/document/media review — instead of guessing values or shipping a static comparison grid. Lookdev = any human tuning of AI output, text and media included, not just sliders.

  • Ci smoke

    ModelBound/modelbound-cli/.github/skills/ci-smoke Skill

    no license1★ repo

    Smoke-test skill for ModelBound CI — triggered when validating GitHub Actions skill checks.

    The same file is in 4 other repositories

  • Decision council

    LeonardSEO/decision-council Skill

    1★

    Use when a user needs a consequential choice, verdict, recommendation, second opinion, or pressure test where multiple criteria or perspectives matter. Use for vendor selection, strategy, hiring, investment, architecture, policy, or any question where a poor decision has material cost. Do not use for factual lookups, simple calculations, rewriting, or low-stakes questions with one obvious answer.

  • New repo

    PBNZ/repo-kit/plugins/repokit/skills/new-repo Skill

    1★ repo

    RepoKit — a reusable repo standard + /new-repo scaffolder, packaged as a Claude Code plugin

  • Npx skill install

    OpenGHz/npx-skill-install/npx-skill-install Skill

    1★ repo

    Install agent skills with `npx skills` the safe, no-duplicate-work way. Before installing, it checks (in one shot) whether the skill is already installed and whether every agent's skills directory is symlinked onto one shared hub — so you install ONCE instead of once per agent. Use this whenever the user wants to add, install, or set up a skill via `npx skills` / the skills CLI, hands over a skills repo or a skill name to install, asks why a skill isn't showing up in some agents, or wants to share one skills folder across Claude Code / Codex / Cursor / Gemini / Copilot / OpenCode. Triggers on: "install this skill", "add the X skill", "npx skills add", "set up this skills repo", "安装这个技能", "装一下这个 skill", "把这个技能装上", "用 npx skills 安装", "让所有 agent 都能用这个技能", "技能装了吗 / 是不是已经安装了", "把各个 agent 的 skills 目录软链接到一起".

  • Gpt image 2 apikey skill

    junxxxxiao/gpt-image-2-apikey-skill Skill

    no license1★

    Generate or edit images through a separate API-key-backed GPT Image 2 provider without using Codex's own OpenAI/Codex credentials. Use when the user asks Codex to create, render, draw, generate, edit, modify, restyle, or transform an image with gpt-image-2/image-2 via an isolated API key stored in a private user config file, especially when the current Codex session is already authenticated with a different API key.

  • Type led launch design

    dvnc-labs/type-led-launch-design Skill

    1★

    Apply a type-led DESIGN.md standard when building or reviewing technical product launch pages, editorial SaaS landing pages, or premium waitlist pages where typography must carry the first viewport. Not for dashboards, generic UI polish, ecommerce catalogs, brand strategy, or broad component libraries.

  • Checkpoint

    vonargo/containuity/skills/checkpoint Skill

    1★ repo

    Use at a clean breakpoint in long-running work — a finished unit of work, before a compaction, or the end of a thread — to preserve continuity across sessions and compactions. Also when the user types /checkpoint, or asks to checkpoint, save progress, snapshot state, or wrap up before context fills.

  • AI SKILL lumberyard

    LevyBytes/AI-SKILL-lumberyard Skill

    1★

    Use when work depends on AWS Lumberyard or CryEngine-family facts: the editor, asset pipeline, Cloud Canvas and AWS gems, and Lua scripting; entities, components, EBus, slices and prefabs; CGA/CAF/CHRPARAMS/Geppetto animation; materials, MTL, shaders, decals and stencils; VFX, particles and rendering; and Wwise/ATL audio. Routes to six area sub-skills. Historical reference (Lumberyard is archived; succeeded by Open 3D Engine).

  • Typo3 site conformance

    netresearch/typo3-site-conformance-skill/skills/typo3-site-conformance Skill

    no license1★ repo

    TYPO3 site/project conformance skill — deployable repos (type:project + Docker/Compose): layout, containers, Concourse CI, supply-chain, secrets, Valkey/ofelia. Sibling to typo3-conformance.

  • Superoffice skills

    cskwork/superoffice-skills Skill

    1★

    Business-document maker for Korean managers and executives. Generates, reads, and converts docx/pptx/xlsx/hwpx/pdf cross-platform with company branding and gate-checked output. Use for /superoffice, weekly/quarterly reports, docx proposals, pptx decks, xlsx settlement sheets or KPI dashboards, hwpx official letters, applying a company template/brand kit, extracting tables, or converting to pdf. Workplace documents, not educational handouts.

  • Jobspipe cli

    jobspipe/jobspipe-cli Skill

    1★

    Search live job postings and detect a company's tech stack with the JobsPipe API. Use when an agent needs real hiring data (open roles, hiring signals, sourcing, salary and location) or to identify which technologies a domain runs.

  • Software design engineering

    benedictdavon/software-design-engineering Skill

    1★

    Use this skill for non-trivial software design, code review, refactoring, debugging, implementation planning, and test strategy where architecture, maintainability, module boundaries, complexity, or engineering tradeoffs matter. Do not use for one-line fixes, formatting-only edits, prose-only tasks, or framework/API documentation lookup.

  • Signal lens

    simhanson123/signal-lens/skills/signal-lens Skill

    1★ repo

    Maintainer PR review for open-source repos: CI weakening, security boundaries, duplicate utilities, test coverage gaps. Auto-selects MCP or CLI. Use before merge, on PRs, or /signal-lens. Not MCP-only — primary maintainer review skill.

  • Agent spawner

    Razaib-khan/ForgeWeave/src/forgeweave/templates/claude/.claude/skills/agent-spawner Skill

    1★ repo

    Creates and configures subagents dynamically with defined roles, constraints, and tool permissions

  • Curl exfil demo

    SuperMarioYL/capsule/examples/skills/curl-exfil-demo Skill

    1★ repo

    A deliberately malicious demo Skill that tries to exfiltrate secrets over the network and read ~/.ssh/id_rsa. Used to show Capsule blocking the calls at the call site.

  • Capture

    nyechiel/ai-augmented-workflow/skills/capture Skill

    1★ repo

    Auto-discover unprocessed meeting notes and transcripts, capture ad-hoc thoughts and facts, or manually capture key facts from a session. Keeps domain knowledge files up to date.

  • Deterministic design skill

    connerkward/deterministic-design-skill Skill

    1★

    Deterministic, measurable design verification — render the UI and PROVE it's balanced and usable instead of trusting the model's eye. Layout audit (centroid / optical-center / pixel-oracle balance via deterministic math, annotated screenshot) + a vision-judged Nielsen usability audit by a SEPARATE fresh-eyes judge. Use to catch AI-generated UI that "looks off", is misaligned / centered-mush, or fails usability — the measurement+judgment layer that token-and-taste design skills (including the default) don't have.

    The same file is in 1 other repository

  • Deepseek reasoning tier clarifier

    vladcheck/skills/deepseek-reasoning-tier-clarifier Skill

    1★ repo

    Analyze task complexity and recommend DeepSeek reasoning tier (low/medium/high/max). Check conversation context and optional codebase files before suggesting a tier. Use when user asks about reasoning effort or when task scope is ambiguous.

  • Vd claim ledger

    jpoindexter/verified-done/skills/vd-claim-ledger Skill

    1★ repo

    Before any completion or success claim, build a claim→evidence ledger. Tag each sub-claim ✅ executed (ran it, saw the output), ◐ code-path (read the code, did not run), or ○ assumed, and state explicitly what each piece of evidence does NOT establish. Only ✅ across every Done criterion counts as done. Use whenever you're about to say something works, is wired, passes, or is complete.

    The same file is in 1 other repository

  • Developer tool installation policy

    gjbex/scientific-computing-skills/skills/developer-tool-installation-policy Skill

    1★ repo

    Decide whether developer and scientific workflow tools belong in project-local environments, pre-commit or CI configuration, or isolated global installs such as pipx, uv tool, or pixi global.

  • Detectoracle

    bzcsk2/detectoracle-skill/skills/detectoracle Skill

    1★ repo

    Scan, mine, and review code using OSS bug-detection patterns. Profile projects, batch-mine GitHub issues, and review local code with evidence.

  • Accuse

    calumjs/sealed-room/plugins/sealed-room/skills/accuse Skill

    1★ repo

    Make a final accusation to solve the murder mystery. Use when the user says "accuse <name>", "I accuse ...", "it was <name>", "my answer is <name>", "the murderer is <name>", or runs /accuse.

  • Sina help

    srahnama/Sina/skills/sina-help Skill

    1★ repo

    Quick-reference card for all Sina modes, skills, and commands. One-shot display, not a persistent mode. Trigger: /sina-help, "sina help", "what sina commands", "how do I use sina".

  • Ask

    DustinVK/wherefore/plugins/wherefore/skills/ask Skill

    1★ repo

    Answer questions about past technical discussions, user stories, and why things were implemented a certain way by searching the team's wherefore log. Use this whenever the user asks why a feature was built the way it was, what was decided about a particular user story, topic, or component, or wants to recall an earlier conversation -- e.g. "why did we...", "what did we decide about...", "is there anything in the log about...", "how were we planning to implement...". Trigger even when the user doesn't name the wherefore explicitly but is clearly asking about a prior decision or its rationale. If nothing relevant is found, say so plainly rather than guessing.

  • Chrome store assets

    JCreatesGH/claude-skills/chrome-store-assets Skill

    1★ repo

    Generate Chrome Web Store (and general app-store) graphics on macOS without any image library — store icon (128×128), small promo tile (440×280), marquee (1400×560), and framed 1280×800 screenshots — by designing in SVG and rasterizing with the built-in `sips` tool. Includes a reference for the CWS listing fields (permission justifications, privacy-practices answers, categories, character limits, and the "Title/Summary are locked to manifest" gotcha). Use when preparing a Chrome extension (or any app) for store submission, or whenever you need exact-dimension PNGs with text/gradients/embedded images on a Mac and ImageMagick/node-canvas are unavailable.

  • Curating readme

    liang-senbei/curating-readme Skill

    no license1★

    Audits a repository's existing information (manifests, code, scripts, config, git history, existing docs) and updates the README plus related documentation (CONTRIBUTING, CHANGELOG, docs/, SECURITY) into a standardized, well-structured form modeled on conventions used by large, reputable open-source projects. Use whenever the user wants to write, update, refresh, standardize, audit, or fix a README or project documentation; mentions "README", "readme", "docs", "文档", "说明文档", "整理文档", "更新 readme"; asks to "document the project", "clean up the docs", or bring documentation in line with conventions. Grounds every statement in the actual repository and never fabricates.

  • Emotion skill

    8903211/emotion-skill Skill

    no license1★

    Analyze authorized WeChat chat exports for self-patterns in relationships, relationship meaning, conversation dynamics, emotional triggers, boundaries, practical reply suggestions, and durable Obsidian/Markdown relationship notes. Use when the user provides WeChat/Weixin chat records or asks to go from exporting/downloading chats, selecting conversations or date ranges, reverse-analyzing who they are in a relationship, understanding what a relationship means to them, producing chat advice, or saving reusable relationship knowledge and phrase frameworks.

  • Ax interface analysis

    Lightbridge-KS/agent-skills/plugins/coding/skills/ax-interface-analysis Skill

    1★ repo

    Analyze the agent experience ("AX") of an interface an AI agent drives — a CLI, MCP server, HTTP/REST API, or library/SDK — against the principles of what makes a system easy for a reasoning-loop agent to use, and prescribe prioritized fixes. Use this skill only when the user explicitly invokes it by name (`ax-interface-analysis`) or near-match mentioning.

  • Asc marketing manager

    superturboryan/app-store-connect-skills/skills/asc-marketing-manager Skill

    1★ repo

    Sync App Store Connect marketing metadata and localized screenshot assets safely, including localized text fields, App Review fields, release-version setup, and screenshot upload/replacement from local folders. Use when updating App Store copy, uploading screenshots, or starting a new editable App Store version.

  • Agent plan mode skill

    wildanfadh/agent-plan-mode-skill Skill

    1★

    Use this skill whenever the user asks for plan mode, planning mode, planning-first workflow, read-only analysis before implementation, or wants an agent/tool that lacks native plan mode to avoid making changes until a plan is approved. This skill is especially important when the user says phrases like "mode plan", "buat rencana dulu", "jangan edit dulu", "analisis dulu", "read-only dulu", "no changes yet", or asks to use an agent while preventing edits. It makes the agent inspect context safely, produce a clear execution plan, and wait for explicit user approval before modifying files, running write operations, committing, pushing, or changing system state.

  • Deep planner

    fengsihao23-eng/shen-hua/skills/deep-planner Skill

    1★ repo

    Shen-Hua deep planning workflow for complex tasks before implementation. Use before large edits, refactors, architecture changes, public API changes, cross-module work, risky migrations, unclear requirements, or when the user asks Codex to investigate first, produce a deep plan, identify risks, or plan a long-running implementation.

  • Skill router

    wpq250dollars/codex-skill-router/skill-router Skill

    1★ repo

    skill selection, goal, plan, high-risk task detection, browser automation, login workflow, config changes, GitHub operations, database work, project scaffolding, multi-skill workflow. Use to decide which Codex skills to use, whether Goal or Plan mode is appropriate, whether read-only inspection is needed, and how to chain multiple skills safely.

  • Analyze

    Colacn/agent-pipeline/skills/analyze Skill

    1★ repo

    Use this skill when the user needs intent clarification, ambiguity resolution, or structured analysis without implementation advice. It asks clarifying questions when uncertain and analyzes only user-provided information after confirmation; use PRD-style packaging only when explicitly requested.

  • Goal

    Jcapathy/loop-goal-skills/goal Skill

    1★ repo

    Objective-mode entry point for the Loop operating cadence. Use when the user wants to define a desired OUTCOME — one or more measurable targets — and have Claude drive the Plan, Build, Test, Reflect, Improve cycle autonomously, logging each run, until every target is verifiably met. Triggers on /goal, "set a goal", "define the objective", "the outcome I want is", "run until it hits", "keep iterating until", "drive this to a number", "don't stop until", or any request that names an end-state plus a success threshold. Pairs with the loop skill: goal captures and sharpens the objective and its definition of done, then runs the loop toward it. Each outcome needs a concrete, evidence-checkable success criterion so TEST can judge pass or fail honestly. Not for vague exploration with no measurable target, and not for one-off single-step tasks that need no iteration.

  • Agent teams

    HorizonBrute/Standardized_AI_Looping_Language-SAILL/skills/agent-teams Skill

    1★ repo

    Create or edit agent-team definitions in local.agent_teams.md at any scope (OS, project, brain, or subfolder) — define the agent chain, each role's model-group preference, and loop/retry constructs. Use when the user types /agent-teams, asks to "add an agent team", "define an agent team", "edit agent teams", or "create a team for this project/scope".

  • Sub2api

    yunhai-dev/sub2api-cli/skills/sub2api Skill

    1★ repo

    Manage a sub2api API gateway instance via the s2a CLI. Use when the user asks to manage users, accounts, groups, subscriptions, settings, proxies, ops monitoring, dashboards, redeem codes, promo codes, channels, backups, or any admin operation on a sub2api instance.

  • Analytics setup

    agent-k-labs/shipping-skills/skills/analytics-setup Skill

    1★ repo

    Set up minimum viable analytics for a vibe-coded app — the right tool stack, the 5 events that matter before 100 users, and copy-paste implementation code. Use when the user wants to add analytics, set up tracking, asks "what should I measure," "I don't have any analytics," "how do I track signups, conversions, or revenue," or has analytics installed but nothing useful being tracked.

  • Project Docs Builder

    m2kky/Project-Docs-Builder Skill

    1★

    Build, review, update, and audit production-grade project documentation packs from a brief, partial docs, or implementation changes. Use when the user asks to create a full docs folder, PRD/FRS/NFR, architecture docs, ERD/data dictionary, API/OpenAPI docs, UX flows, project plans, engineering workflow docs, AI agent runbooks, operational runbooks, or a documentation system similar to a mature product/architecture repository.

  • Hello world

    jnMetaCode/skillet/examples/skills/hello-world Skill

    1★ repo

    A tiny example skill that greets the user — use it to verify skillet installs work.

  • Voice

    segudev/voice-skill/skills/voice Skill

    no license1★ repo

    Speak the agent's reply aloud through Kyutai Pocket TTS. Use whenever the user asks to hear the answer instead of (or in addition to) reading it - triggers like "use voice to respond", "voice this", "say that aloud", "read your reply out loud", "dictate the answer", "respond with voice", or "turn voice mode on/off".

  • Skill forge

    truthandstories07-sudo/skill-forge Skill

    1★

    Create, evaluate, and improve agent skills (SKILL.md format) using an evidence-based methodology distilled from 244 top skills across the Anthropic, OpenAI, vendor, and community ecosystems. Use when the user wants to turn a repeated task, workflow, or expertise into a skill, asks to create or scaffold a SKILL.md, wants an existing skill judged, scored, or audited, or asks to improve a skill's structure, description, or triggering (中文触发:做成技能、创建 skill、沉淀为 skill、写一个技能、评估/评判这个 skill、优化 skill 描述、skill 体检). For routing between already-installed skills use agent-skill-hub instead; for Lark API skills, lark-skill-maker supplies the domain template while this skill governs quality and testing. NOT for writing ordinary prompts, CLAUDE.md/AGENTS.md rule files, or subagent definitions.

  • Visual literacy

    amberhoward78-creator/visual-literacy Skill

    no license1★

    Visual Literacy is a reasoning framework that teaches AI to make visual design decisions by modelling human perceptual psychology and expert design heuristics, rather than relying on formatting instructions alone. Use when designing documents, interfaces, presentations, reports, or any visual communication where the goal is comprehension, not just display.

  • Travel itinerary book

    jechiu16/travel-itinerary-book Skill

    1★

    Build fact-checked, mobile-first travel itineraries and multi-day destination guides as shareable single-file HTML handbooks, with optional all-expanded print/PDF output. Use when the user gives trip notes, asks to combine draft itineraries, wants a route proposed from scratch, or needs current hours/prices/closures verified. Triggers on "行程", "旅遊手冊", "itinerary", "trip plan", "booking checklist", or "find the latest info and make it nice".

  • AI SKILL ableton live

    LevyBytes/AI-SKILL-ableton-live Skill

    1★

    Use when working with Ableton Live 12 (the DAW): the Arrangement and Session views; clips and the Clip View; audio warping and tempo; MIDI and MPE editing; MIDI tools; converting audio to MIDI; comping; stem separation; recording; routing and I/O; mixing; automation and clip envelopes; instrument/drum/effect racks; the browser and managing files/sets; Live's built-in instruments, audio effects, and MIDI effects; Max for Live; Push 1 and Push 2; Link/tempo sync; tuning systems; keyboard shortcuts; and accessibility. The official Live 12 Reference Manual. Not for other DAWs.

  • Workflow create

    manansuneja/skills/skills/agent-workflows/workflow-create Skill

    1★ repo

    Connects multiple skills into a reusable workflow. It creates a coordinator that runs individual skills in sequence while keeping every skill usable on its own. It can also create missing steps, adopt skills into one family, or update and remove workflows it created. Use when the user invokes /workflow-create, says "/workflow-create connect these skills", provides skill names or GitHub URLs to chain together, or asks to create, connect, compose, change, repair, validate, or delete a skill workflow.

  • Product page designer

    johnsonbuilds/awesome-hermes-skills/creative/product-page-designer Skill

    no license1★ repo

    Design and generate international-brand-quality product detail pages for cross-border e-commerce. Generates complete product page content (Hero, Features, Lifestyle, Specs, FAQ) plus AI image prompts and visual assets for any consumer electronics product. Output meets Apple/Nothing brand standards.

  • Gh comment image upload

    AppLoidx/gh-comment-image-upload Skill

    1★

    Use when you need to put a local image or screenshot into a GitHub PR or issue comment automatically, with no human drag-drop — especially in a PRIVATE repo, where token-based user-attachments upload fails (422 anti-forgery) and committing to a branch pollutes the diff or triggers CI/Vercel builds. Triggers "attach screenshot to the PR", "embed image in PR comment", "put the image in the comment".

  • Organize jellyfin media

    j0m0k0/jellyfin-skills/skills/organize-jellyfin-media Skill

    1★ repo

    Organize movie and TV/serial folders for Jellyfin naming conventions with IMDb metadata provider IDs. Use when Codex needs to inspect, rename, or restructure a media folder for Jellyfin Movies or Shows libraries, including skipping already-canonical folders, adding `[imdbid-tt...]` identifiers, planning safe renames, and moving sidecars/extras without deleting media.

  • Ai longform reading coach

    Juliejue/claude-longform-reading-skills/ai-longform-reading-coach Skill

    1★ repo

    AI 前沿科技人文精读教练。Use when the user asks to deeply read, interpret, or replace reading of AI longform essays, newsletters, papers, Every, Stratechery, Anthropic Papers, product philosophy, human-AI interaction, cognition, mental models, or AI humanities analysis. Must produce exhaustive Chinese deep reading with bilingual highlights, collection curation first, active batching, and actionable cognitive/workflow guidance.

  • Acceptance criteria

    celestialdust/achilles-skills/skills/acceptance-criteria Skill

    1★ repo

    Turn a prd.md into acceptance.md — the Given/When/Then prose contract that is the SOLE human-anchored oracle for the entire autonomous run. Reach for this the MOMENT a prd.md exists and BEFORE any planning, TDD, or QA — test-driven-development and quality-verification REFUSE to run without a signed acceptance.md. If you are about to write "acceptance criteria", "definition of done", test scenarios, or Given/When/Then for a feature, you need this first. Behavioral-only: keep ALL design floors out (those belong to frontend-design's contract).