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/sepahead/crebain/agents-mdgit clone --depth 1 https://github.com/sepahead/crebainWhat 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 | $0.02950 | $0.02950 |
| Opus 5 | $0.01475 | $0.01475 |
| Sonnet 5 | $0.00590 | $0.00590 |
| Haiku 4.5 | $0.00295 | $0.00295 |
Grade A, and why
crebain 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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 224 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CREBAIN development guide
Build and validation commands
# Frontend development
bun run dev # Start Vite dev server
bun run build # Typecheck + build for production
bun run typecheck # TypeScript type checking only
# Tauri (full app)
bun run tauri:dev # Development mode with hot reload
bun run tauri:build # Production build
# Validation and testing
bun run lint # ESLint
bun run format # Prettier (write); format:check verifies
bun run test # Run tests in watch mode
bun run test:run # Run tests once
bun run test:coverage # Run tests with coverage (enforces thresholds)
bun run test:responsive # Production-browser layout smoke at supported viewport/scale boundaries
bun run benchmark:native-detector -- --help # Release-only native detector evidence CLI
bun run check:bundle # Build + initial-bundle size budget
bun run check:docs-visuals # Verify tracked Markdown diagram coverage and SVG accessibility
bun run check:ncp-coherence # Verify NCP manifests, locks, and normative docs agree
bun run check:ncp-headless-boundary # Verify the isolated headless package boundary
bun run check:phase0-baseline # Verify the frozen Phase 0 command-surface baseline
bun run check:product-profiles # Verify all eight immutable 0.9 NoAuthority profiles
bun run check:ipc-contracts # Verify frontend/Rust commands and event contracts
bun run check:release-tools # Verify version/tag and digest-manifest tooling
bun run check:vendor-compat # Verify exact crates.io overlay provenance
bun run check:production-vendors # Verify pinned Spark/Rapier/Three transforms and local-byte runtimes
bun run check:production-boundary # Production vendors + authority-boundary checks
bun run check:managed-simulation-boundary # Verify the simulator-only Host API 2.0 crate, manifest, schemas, and transcript
bun run check:managed-simulation-inputs # Verify tracked one-, two-, and three-drone real-NEST inputs
bun run check:managed-simulation-contract # Build and test the private-pipe and installed-package contracts
bun run check:managed-simulation-rust # Check the release-profile managed simulation crate
bun run check:ros-defs # Validate ROS definitions and package XML
bun run check:nix-deps # Verify bun.nix is exactly generated from bun.lock
bun run check:plant-boundary # Verify the inert plant package/process dependency boundary
bun run check:plant-frames # Verify the digest-bound JS/Rust frame-convention corpus
bun run check:plant # Check the headless plant-authority package
bun run test:plant # Test command/health/captured-age/safe-action/deadline-monitor/apply-observation contracts plus frame/lifecycle/channel/passive-expiry/headless foundations
bun run test:managed-simulation # Test the release-profile one-to-three-drone core and managed private-pipe adapter
bun run clippy:plant # Strict Clippy for all plant targets
bun run clippy:managed-simulation # Strict release-profile Clippy for the managed simulation crate
bun run doc:managed-simulation # Build managed simulation Rust documentation with warnings denied
bun run fmt:plant:check # Rustfmt check scoped to the plant package
bun run fmt:managed-simulation:check # Rustfmt check scoped to the managed simulation crate
bun run self-check:plant # Run crebain-plantd in inert self-check mode
bun run validate # contracts/provenance + typecheck/lint/format/frontend tests
bun run validate:all # NCP + frontend + inert plant + Rust default/NCP gates
# Rust backend
bun run check:rust # locked cargo check for src-tauri/Cargo.toml
bun run test:rust # locked cargo test for all default targets
bun run clippy:rust # locked cargo clippy for all default targets; warnings denied
bun run fmt:rust:check # Rustfmt check for src-tauri (part of validate:all)
bun run check:rust:ncp # locked check of dormant bridge and Galadriel producer
bun run clippy:rust:ncp # locked clippy for bridge/producer targets; warnings denied
bun run test:rust:ncp # locked tests for bridge/producer targets
bun run check:ncp-headless # locked check of all isolated headless NCP targets
bun run clippy:ncp-headless # strict Clippy for all isolated headless NCP targets
bun run test:ncp-headless # locked tests for all isolated headless NCP targets
bun run self-check:ncp-headless # Run the network-free headless invariant check
cargo build --locked --manifest-path src-tauri/Cargo.toml
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.
- 3d ago First seen · 224 lines · 2,950 tokens per session scan A 6a7f1c89ff7e
crebain AGENTS.md is an instructions file published in the GitHub repository sepahead/crebain (21 stars, last pushed 5d ago), licensed Apache-2.0. It adds 2,950 tokens to every session, about $0.0147 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
ai-engineering-from-scratch AGENTS.md
Instructions for rohitg00/ai-engineering-from-scratch, covering agents.md, philosophy, repo layout, hard rules and dependencies.
FluidAudio copilot-instructions.md
Copilot instructions for FluidInference/FluidAudio, covering fluidaudio - agent development guide, build & test commands, architecture, critical rules and code style (swift-format config).
pixeltable CLAUDE.md
Claude Code instructions for pixeltable/pixeltable, covering claude.md, protected configuration, project overview, repository structure and setup commands.
pixeltable copilot-instructions.md
Copilot instructions for pixeltable/pixeltable, covering expression dsl — most common bug source, type safety, security, code style (don't contradict) and testing.
speech-swift AGENTS.md
Instructions for soniqo/speech-swift, covering agent instructions, communication style, workflow, running tests and benchmarks — sequential and memory-aware and git conventions.
watch-skill AGENTS.md
Instructions for oxbshw/watch-skill, covering working on watch skill, what this project is, layout, commands and rules that are not negotiable.