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/rafmacalaba/armada/agents-mdgit clone --depth 1 https://github.com/rafmacalaba/armadaWhat 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 | $0.01423 | $0.01423 |
| Opus 5 | $0.00711 | $0.00711 |
| Sonnet 5 | $0.00285 | $0.00285 |
| Haiku 4.5 | $0.00142 | $0.00142 |
Grade A, and why
armada 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 — 154 lines — stays where its author put it; the contents beside it link to each section on GitHub.
armada — Build Rules
These rules apply to every agent working on this project. Generated by armada.
The job
Build exactly as specified in armada/REQUIREMENTS.md. That document is the contract: its phases, success criteria and final criteria decide when work is done. When in doubt, armada/REQUIREMENTS.md wins.
The team
- orchestrator — Delivery lead / scheduler. Model
opencode-go/minimax-m3. - backend-dev — Backend implementation. Model
opencode-go/deepseek-v4-pro. - frontend-dev — Frontend implementation. Model
opencode-go/minimax-m3. - qa — Quality assurance. Model
opencode/mimo-v2.5-free. - adversary — Adversarial reviewer. Model
opencode-go/deepseek-v4-pro. - security — Security auditor. Model
opencode/big-pickle. - docs — Technical writer. Model
opencode/deepseek-v4-flash-free. - architect — Architecture / code review. Model
opencode/big-pickle.
Role boundaries are enforced by permissions and are absolute. Do not work around them with shell commands: if the edit tool would deny a file, do not modify that file any other way.
Repository conventions
- End-to-end tests live under
armada/e2e/{feature}/. Only qa writes there. - Screenshots live under
armada/screenshots/{feature}/. - No emojis in code, comments, print statements or logging.
- Keep it simple: small modules, clear names, no defensive programming, no overengineering.
- Prefer popular, well-supported libraries over custom code.
- Shipnames title format: When invoking
task, setdescriptionto the work-only title. Do not manually prefix<Ship>or[role]; the shipnames plugin auto-prefixes them. - Evidence-based receipts: Subagent tasks require pasted evidence in the receipt
Evidenceblock. No claim of done is accepted without evidence.
Adaptive delivery
- Commodore infers risk; QA always active; evidence depth follows risk. Other roles stay standby.
- Ask user risk override only for ambiguity, high consequence, contract conflict, or downgrade.
- Dependency-ready phases and separate voyages run parallel; serialize shared writers. Group findings before fixes; only
BLOCKINGstops work. Use compact receipts.
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 · 154 lines · 1,423 tokens per session scan A f3f3911f8736
armada AGENTS.md is an instructions file published in the GitHub repository rafmacalaba/armada (91 stars, last pushed 4d ago), licensed MIT. It adds 1,423 tokens to every session, about $0.0071 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
agent-orchestrator AGENTS.md
AGENTS.md instructions for Untrivial-ai/agent-orchestrator, covering agents.md, repo layout, commands, where to look first and distribution.
agent-orchestrator CLAUDE.md
Claude Code instructions for Untrivial-ai/agent-orchestrator, covering claude.md, app state lives under /.ao only and design system.
agent-of-empires AGENTS.md
AGENTS.md instructions for agent-of-empires/agent-of-empires, covering repository guidelines, where to look, commands, code rules and tests.
superpowers-zh GEMINI.md
Instructions for jnMetaCode/superpowers-zh, covering superpowers-zh 中文增强版, 核心规则, 可用 skills and 如何使用.
teaql-agent-kit AGENTS.md
AGENTS.md instructions for teaql/teaql-agent-kit, covering agents.md — teaql agent kit and hard requirements.
untether CLAUDE.md
Instructions for littlebearapps/untether, covering untether, features (vs upstream takopi), architecture, key files and reference docs.