CLAUDE

A definition file describing the agents and coordination rules used by the MAKER framework. It assigns roles such as breaking tasks down, editing files, checking results, and reviewing the final work.

In plain words
What is it for?
It is for configuring MAKER agents, including decomposition, execution, validation, review, checkpointing, caching, complexity assessment, and parallel planning.
Why use it?
It makes each agent's responsibility and available tools clear, helping the framework coordinate specialized steps safely.

Agent

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 agents/forsonny/maker-framework/claude
Clone the repo
git clone --depth 1 https://github.com/forsonny/maker-framework
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 503 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00000 $0.00503
Opus 5 $0.00000 $0.00251
Sonnet 5 $0.00000 $0.00101
Haiku 4.5 $0.00000 $0.00050

Measured 2d ago against content hash 065f538cafac, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

CLAUDE 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.

agents/CLAUDE.md · 61 lines

How it starts

The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MAKER Agents - Coordination Rules

This file is lazy-loaded when working with agent definitions.

Agent Responsibilities

Agent Role Model
maker-decomposition Break tasks into atomic steps (m=1) inherit
maker-decomposition-discriminator Validate decomposition quality inherit
maker-solver Execute exactly ONE atomic step inherit
maker-red-flag Validate solver outputs (<750 tokens) haiku
maker-solution-discriminator Vote on critical step candidates inherit
maker-reviewer Final quality gate (zero issues) inherit
maker-complexity-estimator Assess task complexity inherit
maker-checkpoint-manager Create/restore file checkpoints haiku
maker-cache-manager Cache decomposition patterns haiku
maker-parallel-coordinator Plan parallel execution waves haiku

Tool Restrictions

Each agent has specific tools available:

  • maker-decomposition: Read, Grep, Glob (NO Write/Edit)
  • maker-solver: Read, Edit, Write, Bash, Grep, Glob (FULL access)
  • maker-red-flag: Read, Grep (validation only)
  • maker-reviewer: Read, Grep, Glob, WebFetch, WebSearch, Bash, QA MCP tools
  • maker-checkpoint-manager: Read, Write, Bash, Glob
  • maker-cache-manager: Read, Grep, Glob, Write

Output Format Requirements

maker-solver Output

Must include:

  • [TASK_ID], [STEP], [STATUS] (SUCCESS/BLOCKED)
  • ACTION TAKEN, OUTPUT STATE
  • VERIFICATION, NEXT STEP INPUT (if SUCCESS)

maker-red-flag Output

  • [VERDICT]: VALID or FLAGGED
  • [VIOLATIONS]: List if FLAGGED

maker-reviewer Output

  • [VERDICT]: APPROVED or NEEDS_FIXES
  • Per-check findings with Required Fix for each issue

Red-Flag Rules

Flag solver output if:

  • Response exceeds 750 tokens
  • Missing required fields
  • Malformed format
  • Status neither SUCCESS nor BLOCKED

Voting Rules (Critical Steps)

  1. Run 3 solver instances in parallel
  2. Collect all candidates
  3. Discriminator selects best using first-to-ahead-by-k
  4. Use winning candidate's OUTPUT STATE for next step

Read the full file on GitHub · 61 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. 2d ago First seen · 61 lines · 0 tokens per session scan A 065f538cafac

Subscribe to this mod's changes

CLAUDE is an agent published in the GitHub repository forsonny/maker-framework (7 stars, last pushed 8mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 503 tokens. 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.