raven CLAUDE.md

raven CLAUDE.md is an instructions file for coding agents from giggsoinc/raven. It costs 3,506 tokens per session, scanned A, original, MIT.

Raven instructions for Claude Code, describing a per-turn discipline process for session loading, prompt routing, and approval before edits. Raven is a tool that organizes how an agent handles coding requests.

In plain words
What is it for?
Starting Raven, loading its memory, classifying each request, checking its gate rules, and presenting a briefing before writing.
Why use it?
They make the agent's routing and preparation visible and require a briefing before guided changes, reducing unstructured edits.

Instructions file

Installs and runs on its own, but its text points at files inside the plugin that ships it — anything it tells you to read at a ${CLAUDE_PLUGIN_ROOT} path is only there once the plugin is installed.

Install

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.

agentmods
npx agentmods add instructions/giggsoinc/raven/claude-md
Clone the repo
git clone --depth 1 https://github.com/giggsoinc/raven

Wrote 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.

agentmods badge for raven CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/giggsoinc/raven/claude-md.svg)](https://agentmods.dev/instructions/giggsoinc/raven/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/giggsoinc/raven/claude-md"><img src="https://agentmods.dev/badge/instructions/giggsoinc/raven/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 3,506 This file is loaded in full into every session.
When invoked 3,506 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.03506 $0.03506
Opus 5 $0.01753 $0.01753
Sonnet 5 $0.00701 $0.00701
Haiku 4.5 $0.00351 $0.00351

Measured 5d ago against content hash 8dd89ea89a44, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

raven 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 5d 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.

CLAUDE.md · 308 lines

How it starts

The opening of the file, as written. The whole thing — 308 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md — Raven Discipline Engine v5.5.6

This file is the contract. Claude reads this on every turn. The rules below are not aspirational — they are how Claude must behave in this project.

First load (memory)

Same UX every IDE: (1) python3 scripts/ops/raven-first.py --boot (fallback: "${CLAUDE_PLUGIN_ROOT}/scripts/ops/raven-first.py" --boot) — rebuild once; (2) if load=1, Read only memory=; (3) python3 scripts/ops/raven-first.py --session-start and print the banner. If they say raven dashboard, run ide-boot.py --open. Honor educate=guided|off. Do not dump OKF at boot.

Every turn: UserPromptSubmit already toasts 🔀 Router · host= · tier → recommend · why · applied=false. Repeat that line if the hook is missing. Never silent routing. Non-Claude hosts: run python3 scripts/ops/raven-first.py --prompt "…" (fallback: python3 "${CLAUDE_PLUGIN_ROOT}/scripts/ops/raven-first.py" --prompt "…") before any other tool.


🛑 PER-TURN DISCIPLINE CONTRACT — Read This First, Every Turn

These steps run on every user message, before generating any response, file read, Bash command, or tool call.

Step 0 — Model toast (before Andie, before tools)

First two lines of every reply, before any work:

🔀 Router · host=… · TIER → recommend … · why: … · applied=…
💰 total-cost=$… last_turn=$… est=$…
session=<the model writing this>

Claude UserPromptSubmit injects 🔀 and 💰 — still repeat Intent: and session= in the reply. The toaster is not a skip. git log / git status / “what shipped” is not outside the router.

Step 0.5 — Intent (plan / debug / direct)

Third line, before tools, exactly one of:

Intent: plan — {why}
Intent: debug — {why}
Intent: direct — {why}
Intent When Then
plan Architecture, design, strategy, tradeoff, new feature shape Andie / raven-plan. Do not free-style a design.
debug Bug, error, stack trace, regression, “not working” (brownfield) andie-jr. Do not diagnose ad hoc.
direct Factual question, routine read, approved “go ahead” execute Answer / execute. Still print 🔀 + session=. Still Andie one-liner.

Read the full file on GitHub · 308 lines

Changes

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.

  1. 5d ago First seen · 308 lines · 3,506 tokens per session scan A 8dd89ea89a44

Subscribe to this mod's changes

raven CLAUDE.md is an instructions file published in the GitHub repository giggsoinc/raven (5 stars, last pushed 4d ago), licensed MIT. It adds 3,506 tokens to every session, about $0.0175 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.