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/devjasha/raunen/agents-mdgit clone --depth 1 https://github.com/devjasha/raunenWhat 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 | $0.00697 | $0.00697 |
| Opus 5 | $0.00349 | $0.00349 |
| Sonnet 5 | $0.00139 | $0.00139 |
| Haiku 4.5 | $0.00070 | $0.00070 |
Grade A, and why
raunen 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 — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
raunen
A terminal agent for local LLMs. One Go binary, no runtime, no server. Go 1.25 or newer; the only dependencies are Bubble Tea and Lip Gloss.
Working here
go build ./... && go test ./...before proposing a change. Both are fast.gofmtis not optional;go vet ./...should stay silent.- No new third-party dependencies without saying why. The binary being small and self-contained is a feature, not an accident.
Layout
main.go— flags, config, wiring, and the one-shot pathinternal/agent— the tool-use loop. Presentation-free: it emits typedEvents and knows nothing about the terminal. Keep it that way.internal/ui— Bubble Tea model. Everything rendered goes through here.internal/acp— Agent Client Protocol over stdio. A transport in front ofagent.Agent: it must never grow behaviour the terminal does not have.internal/provider— OpenAI-compatible streaming clientinternal/tools— the built-in tools, rooted at the working directory. Every result is bounded at the registry, not in the tool: what does not fit is kept in the result store and reachable through theresulthandle, so an MCP tool cannot bypass the cap by being added laterinternal/attach— loading images from a path or the clipboard, and the size and format limits. The wire form lives inprovider.toWire, not on the message type: a session is saved asprovider.Messagetoo, and the endpoint's shape is not what belongs on diskinternal/fileset— what git considers part of the project; shared by the search tools and by@completion, so the two cannot disagreeinternal/permission— standing allow/deny rules. Gating lives in one place,Agent.dispatch; a tool must never decide its own permissionsinternal/skills— SKILL.md discovery, including other agents' directoriesinternal/mcp— MCP client: stdio, HTTP and SSE transports, OAuth 2.1internal/config— config, skills and MCP definitions on diskweb/— the marketing and docs site, deployed separately. Not the agent.
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 · 56 lines · 697 tokens per session scan A 3588bfd6c1e1
raunen AGENTS.md is an instructions file published in the GitHub repository devjasha/raunen (5 stars, last pushed 11d ago), licensed MIT. It adds 697 tokens to every session, about $0.0035 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
aider-desk AGENTS.md
AGENTS.md instructions for hotovo/aider-desk, covering agents.md, common commands, development, type checking and linting and formatting.
Meta_Kim CLAUDE.md
Instructions for KimYx0207/Meta_Kim, covering metakim for claude code, fast read, what this repository is, claude code's role and canonical vs runtime files.
crystal CLAUDE.md
Instructions for stravu/crystal, covering crystal - multi-session claude code manager, project overview, references, implementation status: ✅ complete and ✅ implemented features.
easy-cheese python.instructions.md
Instructions for paulnsorensen/easy-cheese, covering hard rules, what to flag and what not to flag.
claude-bridge AGENTS.md
Instructions for kadaba/claude-bridge, covering agents.md — using claude-bridge from any ai coding cli, two independent channels, 1. coordination (the conversation), 2. file transfer (the bytes) — same for every tool and workflows.
AmbyKit CLAUDE.md
Instructions for ambystechcom/AmbyKit, covering claude code — notes for the ambykit repo and claude-specific.