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/smallcloudai/refact/agents-mdgit clone --depth 1 https://github.com/smallcloudai/refactWhat 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 | $0.02346 | $0.02346 |
| Opus 5 | $0.01173 | $0.01173 |
| Sonnet 5 | $0.00469 | $0.00469 |
| Haiku 4.5 | $0.00235 | $0.00235 |
Grade A, and why
refact 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 yesterday.
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 — 164 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Refact Monorepo
AI coding assistant: Rust engine (LSP/HTTP server) + React chat UI + IDE plugins (VSCode, JetBrains) + cloud backend.
Repository Map
| Subproject | Path | Language | AGENTS.md |
|---|---|---|---|
| Agent Engine | refact-agent/engine/ |
Rust 2021, async/tokio | ✅ refact-agent/engine/AGENTS.md |
| Agent GUI | refact-agent/gui/ |
TypeScript/React 18 | ✅ refact-agent/gui/AGENTS.md |
| VSCode Extension | plugins/vscode/ |
TypeScript | — |
| JetBrains Plugin | plugins/intellij/ |
Kotlin, Gradle | — |
| Documentation | docs/ |
Astro (static site) | — |
| IDE metadata | .idea/ |
IntelliJ project config | keep local/editor files out of commits; .idea/workspace.xml is ignored |
| Agent notes | .agents/ |
onboarding notes | checked for repo-specific guidance when present |
| Codex workspace | .codex/ |
Codex config/data | checked for repo-specific guidance when present |
Root .gitignore |
.gitignore |
repository ignore rules | includes local editor and build output exclusions; check before adding new generated files |
Sub-project AGENTS.md files contain detailed architecture, patterns, and checklists. Read them before working in those directories.
Verification Commands
Always verify your changes compile and pass tests before finishing. Both engine and GUI builds are heavy — plan accordingly.
Engine (refact-agent/engine/)
cd refact-agent/engine
# Fast check — type/borrow errors only (~1-3 min, no codegen)
cargo check
# Unit + doc tests (~3-8 min first build, ~1-3 min incremental)
cargo test --lib && cargo test --doc
# Full release build (~10-20 min cold, ~2-5 min incremental)
# LTO + opt-level=z + strip — very slow from scratch
cargo build --release
⚠️ First build compiles ~85 crates + 7 tree-sitter parsers + SQLite. Expect 10-20 minutes cold. Incremental builds are much faster. CI runs cargo test --release on 7 platform targets.
Python integration tests (tests/*.py) require a running refact-lsp instance — don't run them as a quick check.
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.
- yesterday First seen · 164 lines · 2,346 tokens per session scan A 1d8fd557364b
refact AGENTS.md is an instructions file published in the GitHub repository smallcloudai/refact (3,540 stars, last pushed 3mo ago), licensed BSD-3-Clause. It adds 2,346 tokens to every session, about $0.0117 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 2026-08-30.
Other instructions, from other repositories
dograh AGENTS.md
Instructions for dograh-hq/dograh, covering dograh - project overview, project structure, tech stack, local development and environment configuration.
commonly CLAUDE.md
Instructions for Team-Commonly/commonly, covering claude.md / agents.md, 🧠 product vision & architecture philosophy, what commonly is, the architecture model and key concepts.
romm CLAUDE.md
Claude Code instructions for rommapp/romm, covering romm - repository guide for contributors & agents, the stack at a glance, deep-dive references, the frontend has two uis - know which you're in and skills - load the focused guide for your task.
romm AGENTS.md
AGENTS.md instructions for rommapp/romm: Read and follow CLAUDE.md for all repository instructions.
agentgram CLAUDE.md
Instructions for agentgram/agentgram, covering claude.md — agentgram development guide, project overview, tech stack, monorepo structure and core architecture patterns.
paperless-brain CLAUDE.md
Instructions for vailsen/paperless-brain, covering claude.md, what this is, running, local dev and app runs at http://0.0.0.0:8080.