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/oath-md/oathmcp/agents-mdgit clone --depth 1 https://github.com/OATH-md/OathMCPWrote 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/instructions/oath-md/oathmcp/agents-md)<a href="https://agentmods.dev/instructions/oath-md/oathmcp/agents-md"><img src="https://agentmods.dev/badge/instructions/oath-md/oathmcp/agents-md.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 | $0.01920 | $0.01920 |
| Opus 5 | $0.00960 | $0.00960 |
| Sonnet 5 | $0.00384 | $0.00384 |
| Haiku 4.5 | $0.00192 | $0.00192 |
Grade A, and why
OathMCP 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 4d 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 — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository Agent Guide
AGENTS.md is the canonical agent guide.
Project Overview
OathMCP is a TypeScript Model Context Protocol (MCP) server implementing 40
established clinical calculators (BMI, GFR, MELD, etc.) as agent-native tools,
with per-value US/SI unit handling, plausibility guards, evidence resources, and
interpreter prompts. It runs over stdio
(npx -y @oath-md/[email protected]), stateless HTTP, or
Cloudflare Workers. The project implements documented instruments; it is not new
clinical research. Read docs/RESPONSIBLE_USE.md before changing public-facing
clinical, deployment, warranty, or responsibility language.
Commands
# Full acceptance check
npm run check
# Individual checks
npm run lint:specs
npm run typecheck
npm run build
npm run test
npm run check:clinical-release
npm run check:package
# Run the MCP server
npm run start:stdio # stdio transport (node dist/server/stdio.js)
npm run start:http # stateless HTTP transport
npx @modelcontextprotocol/inspector node dist/server/stdio.js # inspect tools/prompts/resources
# Exercise one calculator through the engine
npx tsx -e "import('./src/engine/run.ts').then(m=>console.log(m.run('bmi',{weight_kg:70,height_cm:170})))"
Architecture
The server is spec-driven: each live calculator has a strict YAML spec, an exactly typed pure TypeScript compute function, and an implementation-assurance dossier. There is no manual per-calculator server wiring — the MCP surface is derived from the live specs at startup.
Layers (all MCP-SDK imports are confined to src/server/** to keep the protocol boundary isolated):
specs/<id>.yaml— the runtime contract: metadata, evidence, inputs, outputs, bands, and warning/cap rules. Runtime specs contain no test fixtures. Validated bySpecSchema(src/engine/spec-schema.ts) at load time.src/engine/— spec loading (load-specs.ts), the unit-conversion table (units.ts, the single conversion source of truth), the band-expression evaluator (bands.ts, a closed hand-written grammar — noeval), the compute registry (registry.ts), typed errors (errors.ts), and the runnerrun(id, rawInputs)(run.ts) that normalizes units, enforces plausibility/hard-limit/cap rules, calls the compute fn, evaluates bands, and returns a normalizedCalcResultenvelope.src/compute/<id>.ts— a pure function over canonical-unit inputs, registered viaregisterCompute('<id>', fn). Receives inputs already normalized to canonical (US) units.round.tsprovides deterministic half-even decimal rounding.src/compute/index.generated.tsis generated from the spec filenames and owns side-effect registration imports.src/server/build-tools.ts—buildServer()derives, per spec, acalculate_<id>tool (input/output schemas from the spec), anevidence_<id>resource, and (if the spec has apromptblock) aninterpret_<id>prompt. Also registers the agent-dispatch toolsfind_calculator,describe_calculator, andcalculate_panel. Entrypoints:stdio.ts,http.ts,worker.ts.
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.
- 4d ago First seen · 128 lines · 1,920 tokens per session scan A a8ea3a8d00b3
OathMCP AGENTS.md is an instructions file published in the GitHub repository OATH-md/OathMCP (11 stars, last pushed 18d ago), licensed Apache-2.0. It adds 1,920 tokens to every session, about $0.0096 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 2026-08-30.
Other instructions, from other repositories
llm-context.py CLAUDE.md
Instructions for cyberchitta/llm-context.py, covering claude.md, working notes (gitignored) and draining the field notes.
openrouter-mcp-multimodal AGENTS.md
AGENTS.md instructions for stabgan/openrouter-mcp-multimodal, covering agent instructions, before you ship, releasing (read this before publishing), short version and version files (must all match package.json).
seekstone CLAUDE.md
Instructions for shaqmughal/seekstone, covering claude.md, what this repo is, commands, the harness itself (run after npm install) and architecture.
Plonk AGENTS.md
Instructions for ostapondo/Plonk, covering agent rules, layout, adding a module, build & verify and code style.
mcp-dockhand copilot-review.instructions.md
Instructions for strausmann/mcp-dockhand: When reviewing pull requests in this repository.
second-brain AGENTS.md
Instructions for LindaHaviv/second-brain, covering instructions for ai coding agents, what to read first, the rules (non-negotiable), common tasks, the sanctioned way and enforcement, not just instructions.