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 instructions/umpire-tools/umpire/agents-mdgit clone --depth 1 https://github.com/umpire-tools/umpireWhat 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 | $0.01343 | $0.01343 |
| Opus 5 | $0.00672 | $0.00672 |
| Sonnet 5 | $0.00269 | $0.00269 |
| Haiku 4.5 | $0.00134 | $0.00134 |
Grade A, and why
umpire AGENTS.md 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 2d 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Umpire Contributor Guide
AGENTS.md is the canonical repo-level instruction file. Keep CLAUDE.md and .cursor/rules/umpire.md symlinked to this file so local tooling stays aligned.
Repo Shape
- Umpire is a declarative field-availability engine for object-shaped state with interdependent options. It is not limited to forms.
- Root workspaces live in
packages/*.docs/is a separate Astro/Starlight app and is not part of the root Yarn workspaces. - Published packages:
core,react,signals,store,zustand,redux,tanstack-store,pinia,vuex,zod,json,reads,write,drizzle,testing,devtools.
Build And Test
Canonical commands and single-package iteration workflows live in CONTRIBUTING.md. Check there first.
Quick reference (root): yarn test, yarn build, yarn typecheck.
Never invoke turbo, bun, or npm directly — use the Yarn wrappers. Tests depend on Yarn workspace resolution and per-package bunfig.toml preloads; bypassing Yarn skips both.
Architecture
@umpire/coreowns pure logic, graph construction, evaluation,umpire(),scorecard(), andchallenge().@umpire/storeis the subscription adapter foundation.redux,pinia,tanstack-store, andvuexnormalize into it;zustandre-exports it because Zustand already fits the contract.@umpire/reactis intentionally thin.@umpire/signalsis the fine-grained reactive adapter.@umpire/json,@umpire/zod,@umpire/reads,@umpire/write,@umpire/drizzle,@umpire/testing, and@umpire/devtoolsare optional helper packages layered on top of core.
Behavior To Preserve
- Satisfaction is presence-based by default: only
nullandundefinedare unsatisfied unless a field overridesisEmpty. - When describing
fair: false, say the field or value isfoul, notunfair. requireschecks both dependency satisfaction and dependency availability.disablesandoneOfinspect source values only, not source availability.- Multiple rules on the same target are ANDed; use
anyOf()for OR behavior. - Disabled fields must report
required: false. play()suggests resets only for fields that became disabled and still hold stale values.challenge()andscorecard()are debugging surfaces, not app-state inputs.
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.
- 2d ago First seen · 88 lines · 1,343 tokens per session scan A d6e539cc58e3
umpire AGENTS.md is an instructions file published in the GitHub repository umpire-tools/umpire (2 stars, last pushed 19d ago), licensed MIT. It adds 1,343 tokens to every session, about $0.0067 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 instructions, from other repositories
formkit CLAUDE.md
Instructions for formkit/formkit, covering formkit, architecture, packages, core (framework-agnostic) and vue integration.
jadwal-sholat AGENTS.md
Instructions for haikelz/jadwal-sholat, covering repository guide, project structure, commands, implementation conventions and seo.
formkind AGENTS.md
Instructions for khanhcamap2020-sudo/formkind, covering agent guide, commands and change rules.
ai CLAUDE.md
Instructions for formio/ai, covering claude.md, project overview, repository info, skills library and iterating on skills.
ai copilot-instructions.md
Instructions for formio/ai: Do not read, reference, or include content from any .env files (except .env.example). These files contain secrets and must never appear in suggestions, completions, or chat responses.
sigwire AGENTS.md
AGENTS.md instructions for yeet-src/sigwire, covering building yeet dashboards, mental model, build bottom-up: data → component → layout, 1. get the data right first, in isolation and 2. build each component against a fake signal.