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/cesarfrancots/xolotl-code/claude-mdgit clone --depth 1 https://github.com/cesarfrancots/Xolotl-CodeWrote 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/cesarfrancots/xolotl-code/claude-md)<a href="https://agentmods.dev/instructions/cesarfrancots/xolotl-code/claude-md"><img src="https://agentmods.dev/badge/instructions/cesarfrancots/xolotl-code/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 | $0.01337 | $0.01337 |
| Opus 5 | $0.00668 | $0.00668 |
| Sonnet 5 | $0.00267 | $0.00267 |
| Haiku 4.5 | $0.00134 | $0.00134 |
Grade A, and why
Xolotl-Code 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 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 — 71 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 this is
Xolotl Code — a desktop coding-agent platform for evaluating, comparing, and racing LLMs on real engineering work. It ships as two surfaces that share one engine and one config file:
- Tauri desktop app (
tauri-app/) — the primary surface: chat pane, agent/team worktree orchestration, an eval lab (incl. Goal Eval), and settings for providers/skills/MCP. xolotlCLI (rust/crates/rusty-claude-cli/) — the same engine without the UI: multi-provider routing, slash commands, sessions, planning, tools, memory.
Both read ~/.xolotl-code/config.json, so a key configured in either tool works in both. (Legacy ~/.claw-code/config.json is auto-migrated on first run.)
Two separate Rust trees
There are two independent Rust build trees — do not confuse them:
rust/— a Cargo workspace (crates/*) that is the shared engine + CLI. Has its own CI (.github/workflows/rust.yml, Rust 1.95.0, Linux+Windows+macOS).tauri-app/src-tauri/— the desktop app's Rust backend (separate crate, not part of therust/workspace). Has its own CI (.github/workflows/tauri-app.yml).
The engine logic lives in rust/crates/runtime/ (conversation loop, prompts, memory, tools, agent supervisor, streaming events). The Tauri backend wraps/reimplements engine concerns for the desktop and exposes them over IPC.
Commands
Rust workspace (rust/)
cargo build --workspace
cargo test --workspace --exclude compat-harness # exclude compat-harness (external checks)
cargo test -p runtime <test_name> # single test
cargo fmt --all -- --check
cargo clippy --workspace --all-features --exclude compat-harness -- -D warnings
cargo install --path crates/rusty-claude-cli --force # install the `xolotl` CLI
Clippy runs with pedantic + -D warnings and unsafe_code = "forbid" (see workspace lints in rust/Cargo.toml) — new code must be clean under both.
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 · 71 lines · 1,337 tokens per session scan A db6abb07d53e
Xolotl-Code CLAUDE.md is an instructions file published in the GitHub repository cesarfrancots/Xolotl-Code (5 stars, last pushed 2mo ago), licensed MIT. It adds 1,337 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
openinterpreter AGENTS.md
Instructions for openinterpreter/openinterpreter, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
amp-acp AGENTS.md
Instructions for tao12345666333/amp-acp, covering agents.md, commands, architecture and code style.
roamcode AGENTS.md
AGENTS.md instructions for burakgon/roamcode, covering repository instructions for coding agents, public-repository safety, stable release and ota contract, releasing a stable version and ota changes.
dev-workspace AGENTS.md
Instructions for iwe-org/dev-workspace, covering agent operating manual, start of every session, the operating loop, conventions and iwe basics.
kumo-coding-agent AGENTS.md
Instructions for kumo-ai/kumo-coding-agent: All instructions, routing tables, and rules are in CLAUDE.md. Read that file as the entry point for this agent.
tidyfactor-design AGENTS.md
Instructions for alwkala/tidyfactor-design, covering ⚡ skill & 24 modular slash commands (7 lifecycle stages) and critical architecture (non-negotiable).