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/zenml-io/kitaru/claude-mdgit clone --depth 1 https://github.com/zenml-io/kitaruWrote 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/zenml-io/kitaru/claude-md)<a href="https://agentmods.dev/instructions/zenml-io/kitaru/claude-md"><img src="https://agentmods.dev/badge/instructions/zenml-io/kitaru/claude-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 5.1 | $0.03380 | $0.03380 |
| Opus 5 | $0.01690 | $0.01690 |
| Sonnet 5 | $0.00676 | $0.00676 |
| Haiku 4.5 | $0.00338 | $0.00338 |
Grade A, and why
kitaru CLAUDE.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 — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What is Kitaru?
Kitaru is a runtime for recording, replaying, and improving AI agents in production. Positioning surfaces (README, docs leads, PyPI, marketing) lead with record → replay → improve (diagnose failures, test model/prompt swaps via replay overrides, compare cost and quality, ship updates with confidence).
Project layout
src/kitaru/ # Python package (src layout) — see src/kitaru/AGENTS.md
analytics/ # Async analytics client and event source tracking
api_models/ # Versioned request/response DTOs shared by server and SDK
client/ # Async SDK making REST calls
server/ # FastAPI server (API, application, domain, adapters layers)
tests/ # pytest tests — see tests/AGENTS.md
examples/ # Runnable SDK examples: python/ (adapter examples) and typescript/ (SDK + adapter examples)
docs/ # Three docs surfaces — see "Documentation surfaces" below
book/ # GitBook source for docs.zenml.io/kitaru (hand-written .md)
content/docs/ # FumaDocs SDK reference content (generated)
scripts/ # Node-side doc generation (convert-sdk-docs.mjs)
app/ # Next.js app routes for the sdkdocs.kitaru.ai reference site
worker/ # Cloudflare worker: redirect.mjs (kitaru.ai/docs) + routing maps
scripts/ # OpenAPI generation, doc generation, and UI bundle scripts
generate_openapi.py # Writes openapi/openapi.json from the FastAPI app
# After regenerating, also regenerate the TypeScript
# types via `pnpm run generate`
# (the typescript CI job fails otherwise)
check_openapi.sh # Fails when openapi/openapi.json is out of date
download-ui.sh # Bundles stable/prerelease Kitaru UI releases into the package tree
FRONTEND-TESTING.md # Read first for Kitaru UI bundle/frontend testing,
# stable/prerelease release validation, and token boundaries
docker/ # Dockerfiles — see docker/CLAUDE.md for full architecture details
design/ # Design docs, meeting notes (gitignored, never commit)
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 Changed · +4 lines · +58 tokens per session df77b836176c
- 6d ago First seen · 159 lines · 3,322 tokens per session scan A 4e2b6732e920
kitaru CLAUDE.md is an instructions file published in the GitHub repository zenml-io/kitaru (279 stars, last pushed yesterday), licensed Apache-2.0. It adds 3,380 tokens to every session, about $0.0169 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
HealthChain CLAUDE.md
Claude Code instructions for healthchainai/HealthChain, covering healthchain - claude code context, 0. project overview, 1. non-negotiable golden rules, 2. testing discipline and 3. build, test & utility commands.
zero-to-ai snyk_rules.instructions.md
Snyk Security At Inception.
noeta AGENTS.md
Instructions for initxy/noeta, covering communication, language, workflow, context / adr and engineering constraints.
rigforge AGENTS.md
AGENTS.md instructions for mrodgersjs-web/rigforge, covering agents.md — rigforge, mission, hard rules, default commands and load order.
noeta CLAUDE.md
Instructions for initxy/noeta, a project described as: The server-side runtime for AI agents. Durable & crash-safe (event-sourced), multi-worker scheduling, per-session sandboxing, human-in-the-loop, provider-neutral. Host agents like Claude Code on your own infra — self-hostable, no vendor lock-in.
workflow AGENTS.md
AGENTS.md instructions for vercel/workflow, covering agent instructions, architecture, core components, workflow execution model and development commands.