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/metio/ilo/claude-mdgit clone --depth 1 https://github.com/metio/iloWhat 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 | $0.02653 | $0.02653 |
| Opus 5 | $0.01326 | $0.01326 |
| Sonnet 5 | $0.00531 | $0.00531 |
| Haiku 4.5 | $0.00265 | $0.00265 |
Grade A, and why
ilo 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 today.
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 — 167 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 ilo is
ilo is a small CLI that manages reproducible build environments by wrapping a
container runtime (Podman, nerdctl, or Docker). ilo shell opens an
(interactive) shell inside a container built from your project's image or
Containerfile; ilo compose does the same backed by a compose file, and
ilo devcontainer / ilo devfile build on those by reading a devcontainer.json
or devfile.yaml and injecting its lifecycle commands. It is a single-binary tool
(picocli + GraalVM native image), written in Java 25.
This repository is the source of ilo itself, and it dogfoods ilo to build
itself — see the dev/ RC files below.
Build & test
This machine has no JDK/Maven/Hugo installed; ilo (a native binary) is on
PATH and everything builds inside the metio/devcontainers-graalvm container.
The dev/ files are ilo argument files invoked with the @ prefix:
ilo @dev/env # interactive bash shell in the GraalVM devcontainer
ilo @dev/build # mvn verify (compile + all tests, skips native build)
ilo @dev/native # mvn verify -Dwith the native executable
ilo @dev/serve # hugo dev server for docs/ at http://localhost:1313
ilo @dev/website # build the docs/ site
The underlying Maven commands (run them inside ilo @dev/env to iterate, e.g.
on a single test):
mvn verify # full verification, native build skipped
mvn verify -Dtest=ShellRuntimeTest # one test class
mvn verify --define skipNativeBuild=false # include the native image
mvn verify is the canonical gate — it runs JUnit 5 tests, the ArchUnit
architecture tests, and JaCoCo coverage. The coverage gate is a 0.75 bundle
instruction ratio plus CLASS MISSEDCOUNT = 0 (every class must be touched):
the ratio is below 1.0 on purpose because OS-gated code (@EnabledOnOs(WINDOWS)
PowerShell paths, @EnabledOnOs(LINUX/MAC) expansion) is uncovered on whichever
single OS the build runs on, so the per-class MISSEDCOUNT = 0 rule plus PIT is
what actually guards test quality. PIT mutation testing (65% threshold, currently
~87%) runs in a dedicated CI job (mutation in verify.yml), not in mvn verify;
run it locally with mvn test-compile org.pitest:pitest-maven:mutationCoverage.
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.
- today First seen · 167 lines · 2,653 tokens per session scan A dfb863900573
ilo CLAUDE.md is an instructions file published in the GitHub repository metio/ilo (21 stars, last pushed yesterday), licensed 0BSD. It adds 2,653 tokens to every session, about $0.0133 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 2026-09-01.
Other instructions, from other repositories
aipod AGENTS.md
Instructions for maurycy/aipod, a project described as: Safe space for AI code agents. Podman wrapper.
agent-box AGENTS.md
Instructions for 0xferrous/agent-box, covering pre commit, image, returns: /nix/store/ -source, wrappers and portal adr.
docker-buildbox copilot-instructions.md
Copilot instructions for ekino/docker-buildbox, covering docker buildbox project - copilot instructions, project overview, project structure, core files and image structure.
docker-buildbox CLAUDE.md
Claude Code instructions for ekino/docker-buildbox, covering claude.md, project overview, common commands, local development setup and setup python environment (requires python 3.11).
Kilntainers CLAUDE.md
Instructions for Kiln-AI/Kilntainers, covering kilntainers: secure agent sandboxes, specs and commands & tools.
aipod CLAUDE.md
Instructions for maurycy/aipod, a project described as: Safe space for AI code agents. Podman wrapper.