ai-control-framework CLAUDE.md

ai-control-framework CLAUDE.md is an instructions file for coding agents from sgharlow/ai-control-framework. It costs 1,577 tokens per session, scanned A, original, MIT.

Repository instructions for an AI-control framework that defines checks, example implementations, manual checklists, and required startup procedures.

In plain words
What is it for?
Use them when working on the framework, its dashboard, MCP server, validation suite, or decision-risk components.
Why use it?
They clarify that the framework contains specifications and examples rather than a complete implementation, and record its verified status and limits.

Instructions file

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/sgharlow/ai-control-framework/claude-md
Clone the repo
git clone --depth 1 https://github.com/sgharlow/ai-control-framework

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 ai-control-framework CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/sgharlow/ai-control-framework/claude-md.svg)](https://agentmods.dev/instructions/sgharlow/ai-control-framework/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/sgharlow/ai-control-framework/claude-md"><img src="https://agentmods.dev/badge/instructions/sgharlow/ai-control-framework/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 1,577 This file is loaded in full into every session.
When invoked 1,577 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 5.1 $0.01577 $0.01577
Opus 5 $0.00788 $0.00788
Sonnet 5 $0.00315 $0.00315
Haiku 4.5 $0.00158 $0.00158

Measured 5d ago against content hash 0197a6712b77, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

ai-control-framework 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 · 185 lines

How it starts

The opening of the file, as written. The whole thing — 185 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE CODE CONFIGURATION

AI Agent Operating Instructions - READ FIRST

FRAMEWORK STATUS (v2.0) — verified 2026-07-18

  • Validation suite: 28/28 passing (validate-framework.sh; in CI)
  • Dashboard tests: 14/14 vitest (in CI, Node 20/22)
  • MCP Server: builds clean (tsc; in CI, Node 20/22). The ad-hoc tests/*.js tool scripts require a live stdio MCP client and are NOT in CI.
  • DRS Components: All 13 implemented with correct weights
  • Retired claim: an earlier "136/136 integration tests" status referenced framework-integration-test.sh, which does not exist in the repo. Do not restate that number.
  • Dogfood DRS: 27/100 self-score (see root README for methodology honesty)

⚠️ CRITICAL: IMPLEMENTATION REQUIRED

This framework uses a "Specification-First, Implementation-Flexible" approach.

The framework provides:

  • SPECIFICATIONS of what to check (ai-framework/specs/)
  • REFERENCE IMPLEMENTATIONS as examples (ai-framework/reference/)
  • MANUAL CHECKLISTS as fallback (ai-framework/reference/checklists/)

YOU MUST:

  1. Read ai-framework/IMPLEMENTATION-GUIDE.md FIRST
  2. Choose or adapt an implementation that works in your environment
  3. Use the specifications to understand WHAT to check
  4. Use the references to understand HOW to check
  5. Fall back to manual checklists if scripts fail

The scripts are NOT magic - they are EXAMPLES that need adaptation!

MANDATORY SESSION START PROCEDURE

  1. Load Framework Files (Every Session)

    • Read ai-framework/templates/code.md - Current session state
    • Read ai-framework/templates/orchestration.md - Control rules
    • Read ai-framework/templates/patterns.md - Implementation patterns
    • Read ai-framework/IMPLEMENTATION-GUIDE.md - How to run checks
    • Perform safety checks per ai-framework/specs/ guidelines
  2. Select Pattern Before Coding

    • NO implementation without pattern selection
    • Match task to PATTERN-XXX from patterns.md
    • Document pattern choice in templates/code.md

Read the full file on GitHub · 185 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 · 185 lines · 1,577 tokens per session scan A 0197a6712b77

Subscribe to this mod's changes

ai-control-framework CLAUDE.md is an instructions file published in the GitHub repository sgharlow/ai-control-framework (4 stars, last pushed 1mo ago), licensed MIT. It adds 1,577 tokens to every session, about $0.0079 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.