What is autoforge CLAUDE.md?

An autonomous software-development framework written in Python that coordinates AI agents through stages such as surveying, implementing, testing, security review, and approval. It can run inside Claude Code, where the subscription provides authentication.

Use it to investigate issues, load SWE-bench tasks, locate relevant code, reproduce failures, apply fixes, run checks, and save patches in the required format.

It organizes broad software tasks into a repeatable process and provides tools for searching, viewing, editing, running commands, and saving patches.

autoforge CLAUDE.md — Instructions file by shigel

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/shigel/autoforge/claude-md

Clone the repo

git clone --depth 1 https://github.com/shigel/autoforge

Per session 713 This file is loaded in full into every session.

When invoked 713 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.

Model	Per session	Once invoked
Fable 5	$0.00713	$0.00713
Opus 5	$0.00357	$0.00357
Sonnet 5	$0.00143	$0.00143
Haiku 4.5	$0.00071	$0.00071

Measured 2d ago against content hash 4ab347c5050e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

autoforge 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 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.

CLAUDE.md · 61 lines

What it actually says

autoforge

自律型AIエージェントによるソフトウェア開発ライフサイクルの完全自動化フレームワーク。

プロジェクト概要

論文: paper/paper.md — アーキテクチャ設計と関連研究
調査: docs/ — OSS調査レポート・競合分析
実装: autoforge/ — Python パッケージ (v0.5)

アーキテクチャ（ハーネスモード）

Claude Code のサブスクリプション内で動作。API key 不要。

Claude Code (推論層 — LLM はサブスクで認証済み)
    │
    ├── autoforge-tools search "pattern"    # コード検索
    ├── autoforge-tools view path           # ファイル閲覧（行番号付き）
    ├── autoforge-tools edit path --old/--new  # 編集（lint検証付き）
    ├── autoforge-tools run-cmd "pytest"    # コマンド実行
    ├── autoforge-tools localize --issue "..." # パターン抽出
    ├── autoforge-tools load-task id        # SWE-benchタスク読み込み
    └── autoforge-tools save-patch id       # パッチ保存（SWE-bench形式）

フルSDLCパイプライン（7エージェント）

Survey → Idea(木探索) → Implement → Test ⟲ → Security(OWASP) ⟲ → Stage → Approval

SWE-bench 競争力スタック

Localizer(3段階) → Reproducer(fail→pass) → ACI(7ツール) → Search Agent → Resolver

ハーネスモードでの Issue 解決手順

タスク読み込み: PYTHONPATH=. python3 -m autoforge.harness.tools load-task INSTANCE_ID --file tasks.jsonl
パターン抽出: PYTHONPATH=. python3 -m autoforge.harness.tools localize --issue "issue text"
コード検索: PYTHONPATH=. python3 -m autoforge.harness.tools search "pattern"
ファイル閲覧: PYTHONPATH=. python3 -m autoforge.harness.tools view path/to/file
修正: PYTHONPATH=. python3 -m autoforge.harness.tools edit path --old "old" --new "new"
テスト: PYTHONPATH=. python3 -m autoforge.harness.tools run-cmd "pytest"
パッチ保存: PYTHONPATH=. python3 -m autoforge.harness.tools save-patch INSTANCE_ID --patch "diff"

開発ルール

Python 3.11+, Pydantic v2, httpx
イミュータブルデータ（frozen dataclass / Pydantic model）
ファイルは400行以下
テスト: PYTHONPATH=. python3 -m pytest tests/ -v

テスト

PYTHONPATH=. python3 -m pytest tests/ -v

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.

2d ago First seen · 61 lines · 713 tokens per session scan A 4ab347c5050e

Subscribe to this mod's changes

autoforge CLAUDE.md is an instructions file published in the GitHub repository shigel/autoforge (5 stars, last pushed 4mo ago), licensed MIT. It adds 713 tokens to every session, about $0.0036 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.

Code review Security Testing

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,182 tokens

buildNext

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

autoforge CLAUDE.md