Menu
[BETA] Start the dev server, open the feature in a browser, and iterate on improvements together.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Install, configure, verify, and refresh required harness runtime readiness facts for spec-first workflows on Claude Code, Codex, Kiro, Qoder, or Cursor.
公开 workflow：编写、改写、迁移或按 audit findings 修复 spec-first source skill（skills/<name>/）时使用。不要用于一次性回答、解释/总结/翻译、只审计、文档导出、第三方安装、普通 spec-* workflow 执行，或手改 generated runtime mirrors。
Use this standalone skill when the user asks to mine a repo's existing coding conventions for future AI coding, generate or refresh project rules with AGENTS.md/CLAUDE.md pointers, create Cursor or Qoder rule files from actual code evidence, or make AI-generated code follow a specific project's habits. Do not use for team standards governance, normal code review/debug/refactor work, linter/formatter configuration, generic best practices, unsupported tool rule files such as .cursorrules or .kiro/steering rules, or generated runtime mirror edits.
Build applications where agents are first-class citizens. Use this skill when designing autonomous agents, creating MCP tools, implementing self-modifying systems, or building apps where features are outcomes achieved by agents operating in a loop.
Capture a visual demo reel (GIF, terminal recording, screenshots) for PR descriptions. Use when shipping UI changes, CLI features, or any work with observable behavior that benefits from visual proof. Also use when asked to add a demo, record a GIF, screenshot a feature, show what changed visually, create a demo reel, capture evidence, add proof to a PR, or create a before/after comparison.
Create, share, view, comment on, edit, and run human-in-the-loop review loops over markdown documents via Proof, the collaborative markdown editor at proofeditor.ai ("Proof editor"). Use when the user wants to render or view a local markdown file in Proof, share markdown to get a URL, iterate collaboratively on a Proof doc, comment on or suggest edits in Proof, HITL a spec/plan/draft for human review, sync a Proof doc back to local, or work from a proofeditor.ai URL. Trigger on phrases like "view this in proof", "share to proof", "iterate with proof", or "HITL this doc", and on spec-brainstorm / spec-ideate / spec-plan handoffs for human review. Also match clear requests for a rendered/shared markdown review surface even if the user does not name Proof. Do not trigger on "proof" meaning evidence, math/logic proof, burden of proof, proof-of-concept, or bare "proofread this" requests where inline text review is expected.
| name | spec-polish-beta |
| description | [BETA] Start the dev server, open the feature in a browser, and iterate on improvements together. |
| disable-model-invocation | true |
| argument-hint | [PR number, branch name, or blank for current branch] |
Start the dev server, open the feature in a browser, and iterate. You use the feature, say what feels off, and fixes happen.
Use when a browser-visible UI exists and the user explicitly wants an interactive polish loop on the current branch or a named PR/branch.
Do not use for non-UI code, backend-only work, headless code review, production deployment, or when the current branch is main/master.
Current branch or PR/branch argument, launch config or detected project type, package manager/port facts, dev-server command, browser URL, and user polish feedback.
Applied UI polish changes, dev-server URL/status, browser observations when requested, and a final commit when the user says the polish pass is done.
Temporary dev-server log files and source changes in the target app; no durable spec-first audit artifact is promised.
Unsafe branch, missing launch/start command, dependency/server startup failure, port conflict, unavailable browser helper, or user stops the loop.
Resolve the branch, start the dev server, open or print the browser URL, iterate on user-visible feedback, verify hot reload or browser state, and commit when done.
Human UI reviewers, spec-code-review, PR preparation, and product/design stakeholders checking the polished experience.
Follows docs/contracts/workflows/scenario-capability-matrix.md (default).
Overrides: none
Resolve all scripts/<name>.sh paths relative to this skill's loaded directory. The target app stays in the user's project checkout; only these helper scripts live beside the skill.
.claude/launch.jsonRun bash scripts/read-launch-json.sh. If it finds a configuration, use it — the user already told us how to start the project.
Run bash scripts/detect-project-type.sh to identify the framework.
Route by type to the matching recipe reference for start command and port defaults:
| Type | Recipe |
|---|---|
rails | references/dev-server-rails.md |
next | references/dev-server-next.md |
vite | references/dev-server-vite.md |
nuxt | references/dev-server-nuxt.md |
astro | references/dev-server-astro.md |
remix | references/dev-server-remix.md |
sveltekit | references/dev-server-sveltekit.md |
procfile | references/dev-server-procfile.md |
unknown | Ask the user how to start the project |
For framework types that need a package manager, run bash scripts/resolve-package-manager.sh and substitute the result into the start command.
Resolve the port with bash scripts/resolve-port.sh --type <type>.
Start the dev server in the background, log output to a temp file. Probe http://localhost:<port> for up to 30 seconds. If it doesn't come up, show the last 20 lines of the log and ask the user what to do.
Load references/ide-detection.md for the env-var probe table. Open the browser using the supported host mechanism (Claude Code browser hint; Codex and terminal contexts print the URL).
Tell the user:
Dev server running on http://localhost:<port>
Browse the feature and tell me what could be better.
This is the core loop. The user browses the feature and tells you what to improve. You fix it. Repeat until they're happy.
agent-browser to screenshot or inspect the page; if it is missing, tell them: "Browser automation helper unavailable. To install/repair, set SPEC_FIRST_BROWSER_HELPER_REQUIRED=1 and rerun spec-mcp-setup (or this host's MCP setup entrypoint). This does not block spec-first baseline." Continue the human browser loop when automated screenshots are unavailable.No checklist. No envelope. Just conversation.
Reference files (loaded on demand):
references/launch-json-schema.md — launch.json schema + per-framework stubsreferences/ide-detection.md — host IDE detection and browser-handoffreferences/dev-server-detection.md — port resolution documentationreferences/dev-server-rails.md — Rails dev-server defaultsreferences/dev-server-next.md — Next.js dev-server defaultsreferences/dev-server-vite.md — Vite dev-server defaultsreferences/dev-server-nuxt.md — Nuxt dev-server defaultsreferences/dev-server-astro.md — Astro dev-server defaultsreferences/dev-server-remix.md — Remix dev-server defaultsreferences/dev-server-sveltekit.md — SvelteKit dev-server defaultsreferences/dev-server-procfile.md — Procfile-based dev-server defaultsScripts (invoked via bash scripts/<name>):
scripts/read-launch-json.sh — launch.json readerscripts/detect-project-type.sh — project-type classifierscripts/resolve-package-manager.sh — lockfile-based package-manager resolverscripts/resolve-port.sh — port resolution cascade