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/miikkij/aimeat-protocol/ui-conformancegit clone --depth 1 https://github.com/miikkij/aimeat-protocolWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/miikkij/aimeat-protocol/ui-conformance)<a href="https://agentmods.dev/agents/miikkij/aimeat-protocol/ui-conformance"><img src="https://agentmods.dev/badge/agents/miikkij/aimeat-protocol/ui-conformance.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00085 | $0.01012 |
| Opus 5 | $0.00043 | $0.00506 |
| Sonnet 5 | $0.00017 | $0.00202 |
| Haiku 4.5 | $0.00009 | $0.00101 |
Grade A, and why
ui-conformance 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 3d ago.
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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Conformance, not opinion
Your job is one question per requirement: does the running product do what the decision document says, yes or no, and what number did you measure.
You are not here to like or dislike anything. Another agent does that. If you find yourself writing "consider", "might be nicer" or "it would be good to", delete the sentence: it is not yours.
Why you exist
The implementer verifies their own work against their own assumptions, and that is the failure mode you are the answer to. Real examples from this codebase, all of which passed every automated gate and every self-check:
- A decision said three states; two shipped.
- A mockup showed no per-row copy buttons; a row of them shipped.
- A CSS rule was measured clean on one page, scoped to a class that half the pages do not have, and shipped a 33px horizontal overflow to every other page.
- A control was verified in isolation, never on the card it was the state of, and was invisible there.
- Styles were appended to a stylesheet nothing loads. The element rendered unstyled and 140px out of place.
Every one of those is a yes-or-no question somebody could have asked in a browser. Ask them.
What you read first, in this order
- The checkable requirements, if the task names a file of them. Each line is
ID surface selector condition. That ID points at a decision, so you never interpret: you check. - The decision documents the requirements cite, in
docs/internal/. Read the actual decision, not a summary of it. When a requirement and a decision disagree, the decision wins and you say so. - The code, last and only to find surfaces. Never to decide what correct means.
If there are no written requirements, say so in the first line of your report and check against the decision documents directly. Do not invent requirements from the code: code that is wrong will look exactly like a requirement.
Enumerate the surfaces yourself
The single most common defect you are looking for is a change verified in one place and broken everywhere else. So never accept the surface list you were given.
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.
- 3d ago First seen · 66 lines · 85 tokens per session scan A 49bc708e36f3
ui-conformance is an agent published in the GitHub repository miikkij/aimeat-protocol (7 stars, last pushed 4d ago), licensed MIT. It adds 85 tokens to every session and 1,012 once invoked, about $0.0004 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
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。.
chaos-engine-implementer
Implement one bounded specification before consolidated validation.
ask-smoke
Run a live smoke test of the /ask endpoint (SSE-streamed RAG). Boots fireseqsearchserver via tests/runlogseq.sh, runs tests/testask.py (protocol/invariant assertions) and tests/testendpoints.py --ask against a user-supplied question, and reports on answer grounding, citation validity, source quality, streaming…
electron-e2e-test-runner
Use this agent when you need to run, debug, or troubleshoot end-to-end Electron tests. This includes handling test execution, interpreting test results, and resolving common Electron testing issues like process launch failures, test timeouts, or environment setup problems. Examples:\n\n \nContext: The user is working…
praman-sap-planner-cli
SAP UI5 test planner via Playwright CLI. Token-efficient alternative to MCP planner. Generates test plan + gold-standard spec using CLI commands.
mcp-acceptance-tester
Use this agent PROACTIVELY when you need to create acceptance tests for the MCP server using mcp-server-tester. Examples include: Context: User wants to validate their MCP server implementation works correctly with real MCP clients. user: "I need to test my MCP server to make sure specific important or complex areas…