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/realactivity/tula/agents-mdgit clone --depth 1 https://github.com/realactivity/tulaWhat 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 | $0.02444 | $0.02444 |
| Opus 5 | $0.01222 | $0.01222 |
| Sonnet 5 | $0.00489 | $0.00489 |
| Haiku 4.5 | $0.00244 | $0.00244 |
Grade A, and why
tula 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 — 234 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md - Tula repo guide for coding agents
You are working in realactivity/tula: the open-source Tula skill and eval repo. Read this file first. It is the canonical orientation for Cursor, Copilot, Codex, Claude Code, and other repo-scoped coding agents.
Not the same file as:
skills/AGENTS.md- how to author OpenClaw skillsdocs/agent/AGENTS.md- runtime workspace playbook on a deployed VM (~/.openclaw/workspace/). Do not copy that file here.
What Tula is (30 seconds)
Tula is a personal health intelligence agent built on OpenClaw. Skills run on a self-hosted VM; health data stays in the user's workspace. This repo ships skills, Waza eval suites, deploy scripts, and the Patient Agent Eval Standard v0.1 (forkable by any patient-facing agent vendor).
Think of Tula as the patient-side equivalent of Microsoft's Scout, built on the same OpenClaw runtime: RealActivity governs the patient agent the way Microsoft governs the workforce agent. Positioning: Microsoft Just Validated Our Architecture.
Watch first:
- YouTube live demo (~16 min) - Epic MyChart at 10:00
- Video guide - both videos, share copy, next steps
- Podcast interview (~17 min)
Fork the eval standard: evals/README.md (78 tasks, 8
skills + composition bundle). Article:
articles/how-will-you-know-if-your-patient-ai-is-working.md.
Repo map
| Path | What lives here |
|---|---|
skills/ |
OpenClaw skills (core product). Each skill: SKILL.md, references/, scripts/ |
evals/ |
Waza eval suites per skill + composition bundle + _templates/ |
docs/ |
Architecture, deploy guide, eval status, build spec, runtime agent templates |
apps/ |
My Aria patient portal UI |
services/wren/ |
Self-hostable SMART on FHIR records relay |
scripts/ |
CI gates, deploy-skills, backup, eval runners |
templates/ |
Tenant onboarding placeholders ({{UPPER_SNAKE}}) |
articles/ |
Published technical writing |
results/ |
Gitignored live eval JSON snapshots (optional publish) |
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 · 234 lines · 2,444 tokens per session scan A 4bdb32036337
tula AGENTS.md is an instructions file published in the GitHub repository realactivity/tula (45 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 2,444 tokens to every session, about $0.0122 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
health-skillz AGENTS.md
AGENTS.md instructions for jmandel/health-skillz, covering agent notes for health-skillz, purpose, configuration discipline, running the server and provider directory assets (/static/brands/...).
ehr-mcp CLAUDE.md
Claude Code instructions for jsfaulkner86/ehr-mcp, covering claude.md — ehr-mcp, what this project is, architecture, module reference and ehrmcp/server.py.
OpenELIS-Global-2 AGENTS.md
Instructions for DIGI-UW/OpenELIS-Global-2, covering agents.md - readme for ai coding agents, file ownership model (014 remediation), project identity, critical prerequisites and java version (mandatory).
breathe-cli CLAUDE.md
Instructions for marekkowalczyk/breathe-cli, covering breathe cli, spec, key constraints, safety constraints (non-negotiable) and testing.
viral-ngs AGENTS.md
Instructions for broadinstitute/viral-ngs, covering agents.md, development environment, docker-centric development, development workflow and running tests.
mosaicatcher-pipeline CLAUDE.md
Claude Code instructions for friendsofstrandseq/mosaicatcher-pipeline, covering claude.md, project overview, version management, common commands and using pixi (recommended for snakemake v9+).