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/goadesign/goa-ai/agents-mdgit clone --depth 1 https://github.com/goadesign/goa-aiWhat 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 | $0.02892 | $0.02892 |
| Opus 5 | $0.01446 | $0.01446 |
| Sonnet 5 | $0.00578 | $0.00578 |
| Haiku 4.5 | $0.00289 | $0.00289 |
Grade A, and why
goa-ai 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 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.
How it starts
The opening of the file, as written. The whole thing — 186 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an agentic systems engineer. Optimize for elegance, strong contracts, conceptual correctness, and less code. Prefer deleting bad abstractions to preserving them. Do not add fallbacks, coercions, or defensive code that hides bugs.
Repository Guidelines
Core Operating Rules
- Plan before acting: for <=2 files, state a brief plan then implement; for >=3 files, write a step-by-step plan first.
- Read before editing. Search instead of guessing.
- Before merging a pull request, address every applicable automated review finding. Verify each finding against the current diff, fix confirmed issues, and explicitly resolve or explain findings that do not require a change.
- Fix root causes, not local workarounds.
- Prefer the simplest design that satisfies the contract. Reduce surface area, delete dead code, and avoid new concepts unless they clearly pay for themselves.
- Be concise in progress updates and summaries.
- Keep one canonical implementation and one source of truth per concept. Delete unused code and commented-out code.
- Before adding an exported type, shared interface, callback shape, wire record, or package dependency, trace the complete producer-to-consumer flow and separate the unavoidable domain concept from implementation plumbing. List which facts the DSL and generator already know, compare private, raw, generated, and public representations, and choose the smallest public contract. Do not export a type merely to avoid an import cycle, simplify one call site, or pass parsed data to code that can privately own decoding. Complete this public-surface and generation-first review before editing; successful tests and a later cleanup pass do not make unnecessary API surface acceptable.
- Validate only at boundaries: HTTP/gRPC handlers, event consumers, DB results, third-party APIs,
ctx.Value(), type assertions, and required map lookups. Inside the codebase, trust Goa and construction-time invariants. - Fail fast on invariant violations. Do not add nil/empty guards, fallback behavior, back-compat fishing logic, or "should not happen" branches for values guaranteed by contracts.
- Do not perform best-effort coercions in runtime/codegen. If a payload, result, or type assertion does not match the contract, return a precise error instead of silently remapping it.
- Configuration belongs in constructors, not environment-variable reads in core logic.
- Keep docs in sync with behavior:
- User-facing
goa-aiDSL, runtime, or codegen changes must updatecontent/en/docs/2-goa-ai/and translated pages when applicable. - Update
README.mdandDESIGN.mdwhen behavior changes.
- User-facing
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 · 186 lines · 2,892 tokens per session scan A c299793285ed
goa-ai AGENTS.md is an instructions file published in the GitHub repository goadesign/goa-ai (8 stars, last pushed 3d ago), licensed MIT. It adds 2,892 tokens to every session, about $0.0145 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.
Other instructions, from other repositories
code-cli AGENTS.md
Instructions for autohandai/code-cli, covering agents.md, project overview, current repository architecture, src/core/agent runtime split (current) and general layout guidance for contributions.
succubus AGENTS.md
Instructions for enowdev/succubus: Several AI agents may work in this repository at the same time. succubus is the shared record of who is here, what the plan is, and which files are taken.
neuraforge-ai AGENTS.md
Instructions for vikisingh23/neuraforge-ai, covering neuraforge ai — agent instructions, how to use, plan before acting, architecture principles (enforce on all code) and code generation.
neuraforge-ai GEMINI.md
Instructions for vikisingh23/neuraforge-ai: Read AGENTS.md for full agent instructions, architecture principles, and code generation/review workflows.
harness-factory CLAUDE.md
Instructions for Kimyongari/harness-factory, covering agent.md, 프로젝트, 명령어 (결정론적 도구를 llm으로 대체하지 말 것), 기계적 강제 (프롬프트 아님, 런타임) and 절대 규칙 (항상 적용).
chatml CLAUDE.md
Claude Code instructions for chatml/chatml, covering claude.md, task management (mandatory), tech stack & directory structure, key entry points and key frontend patterns.