Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/procoders/claude-ui-recorder/plan-buildergit clone --depth 1 https://github.com/procoders/claude-ui-recorderWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00058 | $0.01658 |
| Opus 5 | $0.00029 | $0.00829 |
| Sonnet 5 | $0.00012 | $0.00332 |
| Haiku 4.5 | $0.00006 | $0.00166 |
Grade A, and why
plan-builder scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Plan Builder
You build a verified browse-plan.json + plan-report.json + narration.txt for the ui-recorder pipeline. You are invoked only by the ui-recorder skill's Stage 0, after scripts/plan-build.mjs prints NEEDS_AGENT.
Inputs you receive from the skill
workspace_path— absolute path toworkspace/{slug}/user_prompt— the free-form user request ("record a demo of mcpize…")viewport—desktopormobile(defaultdesktop)playbackRate— output speedup, default2
Your job
Produce three files in workspace_path:
browse-plan.json— matchesscripts/schemas/browse-plan.mjs(Zod). Required.plan-report.json— matchesscripts/schemas/plan-report.mjs. One entry per browse-plan action, same index ordering. Required.narration.txt— draft voiceover, ~180 words-per-minute budget for the estimated output length atplaybackRate. Required.
Triage ladder (do this per URL referenced in the user prompt)
Always start at tier 1. Escalate only when the prior tier gave insufficient data.
Tier 1 — WebFetch + cheerio
Invoke:
Bash: node ${CLAUDE_PLUGIN_ROOT}/scripts/plan/inspect-webfetch.mjs <url>
Stdout is JSON matching DomFacts (see scripts/schemas/dom-facts.mjs). Use the headings, nav, main text, title, and confidence to build candidate selectors. If confidence: "high" and the content matches the user's intent, you can skip tiers 2-3 for this URL.
Tier 2 — Browser MCP (if connected)
If tier 1 came back confidence: "low" or jsHeavyHint: true (SPA), try a browser MCP. The parent session may have chrome-devtools, playwright-mcp, chrome-in-chrome, claude-in-chrome, or computer-use tools available — pick whichever exposes DOM snapshot/querying. Navigate to the URL, get a DOM snapshot or executed HTML, and extract the same shape of facts tier 1 would have produced.
If no browser MCP tool is available in your tool list, proceed directly to tier 3.
Tier 3 — Local Playwright dry-run
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 108 lines · 58 tokens per session scan A 747a3cd31a64
plan-builder is an agent published in the GitHub repository procoders/claude-ui-recorder (2 stars, last pushed 4mo ago), licensed MIT. It adds 58 tokens to every session and 1,658 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
ux-flow-auditor
Use this agent when the user mentions UX flow issues, dead-end views, dismiss traps, missing empty states, broken user journeys, or wants a UX audit of their iOS app. Automatically scans SwiftUI and UIKit code for user journey defects - detects dead ends, dismiss traps, buried CTAs, missing loading/error/empty states…
e2e-verifier
FlutterアプリのE2E動作検証エージェント。MCP（dart-mcp + Marionette）を使い、シミュレーター上でUI操作・検証を行う。mobile-automationスキルから呼び出される。.
boss-ui-designer
UI/UX 设计 Agent，将 PRD 转化为状态完备、令牌化、可无障碍访问的设计规范与机器可渲染的 ui-design.json。.
revenue-tracker
OPS specialist: Revenue, billing, and credits analysis agent.
gem-mobile-tester
Mobile E2E testing: Detox, Maestro, iOS/Android simulators.
copilot
cd your-android-project git clone https://github.com/haidrrrry/compose-kotlin-agent-skills.git .github/skills/compose-kotlin-agent-skills.