agent-rules-skill-auth…
Maintain WebDev Agent Kit source rules, skills, metadata, and validators. Excludes upgrading installed copies.
Onboard frontend projects with local context, client and verification facts. Codex onboarding includes economical subagent setup and activation checks. Excludes model changes during updates, facts-only requests and Plan Mode; no app code or instruction replacement.
$ npx -y skills add ytvee-dev/webdev-agent-kit --skill project-onboarding-adapter --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/project-onboarding-adapterContext preview
The summary Claude sees to decide when to auto-load this skill.
Onboard frontend projects with local context, client and verification facts. Codex onboarding includes economical subagent setup and activation checks. Excludes model changes during updates, facts-only requests and Plan Mode; no app code or instruction replacement.
name: project-onboarding-adapter
description: 'Onboard frontend projects with local context, client and verification facts. Codex onboarding includes economical subagent setup and activation checks. Excludes model changes during updates, facts-only requests and Plan Mode; no app code or instruction replacement.'
id: 'agents.skills.project-onboarding-adapter.skill'
title: 'Project Onboarding Adapter'
doc_type: 'skill'
layer: 'skill'
status: 'active'
publishable: true
local_only: false
skill: 'project-onboarding-adapter'
tags:
- 'agents/skill-package'
- 'agents/onboarding'
- 'frontend/project-context'
parent: []
related:
- '[[templates/project/model-routing-profile]]'
- '[[skills/project-onboarding-adapter/references/codex-model-bootstrap]]'
- '[[skills/project-onboarding-adapter/references/model-workload-matrix]]'
- '[[common/host-instruction-migration-rules|Host Instruction Migration]]'
- '[[common/project-fact-provenance-rules|Project Fact Provenance]]'
- '[[templates/project/verification-profile|Verification Profile Template]]'
- '[[common/core/runtime-core-policy|Portable Runtime Core Policy]]'
- '[[common/readme-policy|README Read And Edit Policy]]'
- '[[profiles/react-typescript/PROFILE|React TypeScript Profile]]'
- '[[skills/project-onboarding-adapter/references/adaptation-checklist|Adaptation Checklist]]'
- '[[skills/project-onboarding-adapter/references/path-audit-checklist|Path Audit Checklist]]'
- '[[common/client-adaptation-policy|Client Adaptation Policy]]'
- '[[common/tool-capability-model|Tool Capability Model]]'
- '[[common/mcp-installation-policy|MCP Installation Policy]]'
- '[[common/target-stack-policy|Target Stack Policy]]'
- '[[common/skill-applicability-policy|Skill Applicability Policy]]'
- '[[common/context-compaction-rules|Context Compaction Rules]]'
- '[[templates/project/client-profile|Client Profile Template]]'
- '[[templates/project/mcp-profile|MCP Profile Template]]'
- '[[skills/project-context-adapter/SKILL|Project Context Adapter]]'
- '[[skills/mcp-toolchain-manager/SKILL|MCP Toolchain Manager]]'
depends_on:
- '[[AGENTS|Canonical Agent Policy]]'Adapt this `.agents` bundle to a host frontend project without creating application source files. For React or Next.js, cache normal target-stack facts. For a non-target frontend project, cache only the detected stack boundary and the framework-agnostic skills that remain usable.
Onboarding must also adapt natively to the installed client target. It creates or refreshes the smallest appropriate host pointer, records client facts in local-only `project/client-profile.md`, and records tool capabilities in local-only `project/mcp-profile.md`.
Route adaptation, initialization, and project-context bootstrap commands here, including `адаптируйся`, `инициализируй .agents`, and `Adapt this .agents bundle to my project.` Full Codex onboarding includes model setup; explicit GPT requests also trigger that phase. A facts-only request does not.
existing project overlays instead of repeating onboarding.
1. Read the host-root native instruction pointer if present and needed for adaptation: `AGENTS.md`, `CLAUDE.md`, or client rules. Inspect the minimal pointer section for detection; read the entire file for requested migration using `common/host-instruction-migration-rules.md`. 2. Read bundle-local `AGENTS.md` when shipped and `common/core/runtime-core-policy.md`. Native plugins use their generated skill prelude; do not require an absent shared-policy entrypoint. 3. Read `common/client-adaptation-policy.md` and only the adapter for the resolved canonical target. 4. Read `profiles/react-typescript/PROFILE.md` and its owning policies only when repository evidence confirms the profile. 5. Read `common/skill-applicability-policy.md` when the detected or suspected stack is outside the target stack. 6. Read `common/tool-capability-model.md` and `tool-capabilities-manifest.json` for capability selection. 7. Read `common/context-compaction-rules.md` when loop memory or resumable workflow setup is requested. 8. Read existing `project/**` overlays when present. 9. Inspect only relevant manifests, configs, source entrypoints, routes, styles, assets, and verification scripts. 10. Read `templates/project/client-profile.md` and `templates/project/mcp-profile.md` before creating those local-only profiles. 11. Read `common/project-fact-provenance-rules.md` and `templates/project/verification-profile.md` before caching verification or capability facts.
Read targeted README sections only when they help identify project intent, setup guidance, or documentation drift. Apply `common/readme-policy.md`, and confirm every cached technical fact through manifests, config, source, CI, package scripts, lockfiles, or real
A cross-platform frontend kit for Codex, Claude Code, Cursor, OpenCode, etc. It helps AI coding agents work predictably and safely with existing frontend projects.
Repo: ytvee-dev/webdev-agent-kit
Maintain WebDev Agent Kit source rules, skills, metadata, and validators. Excludes upgrading installed copies.
Inspect Figma links through MCP or browser computer use, or analyze screenshots and selected-layer properties. Produce frontend specs and clarify unresolved…
Create, resume, analyze, or converge standard and deep frontend execution plans with traceable slices, evidence, and stop/resume state. Skip lightweight…
Plan architecture and ownership for standard or deep React/Next.js work across routing, state, data, styling, forms, builds, migrations, and implementation…
Diagnose and fix frontend defects from evidence, including unclear UI/runtime errors, broken routes, styling regressions, and hydration or client issues.…
Set subject-grounded visual direction for standard or deep UI work, redesign, polish, critique, anti-template review, or design handoff. Skip technical…