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/strale-io/strale/agents-mdgit clone --depth 1 https://github.com/strale-io/straleWhat 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 | $0.03866 | $0.03866 |
| Opus 5 | $0.01933 | $0.01933 |
| Sonnet 5 | $0.00773 | $0.00773 |
| Haiku 4.5 | $0.00387 | $0.00387 |
Grade C, and why
strale AGENTS.md scanned grade C with 1 finding 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
checkout's `node_modules`, then later removing that worktree with `rm -rf`, How it starts
The opening of the file, as written. The whole thing — 264 lines — stays where its author put it; the contents beside it link to each section on GitHub.
About this file
AGENTS.md is the Codex-CLI counterpart to CLAUDE.md. It is a condensed
derivative — CLAUDE.md is canon. Where a fact can drift (capability/solution
counts, decision full text, which solutions are active, exact protocol steps),
this file points at the CLAUDE.md section or the canonical source (the
solutions/capabilities tables, GET /v1/platform/facts) instead of
restating it. Refresh this file against CLAUDE.md when drift is noticed —
don't let it silently go stale again. (It was ~4 months stale as of
2026-08-17; see docs/company/CHARTER.md and the Report Filing Convention
section of CLAUDE.md for the history.)
Session Start
- Declare session intent (one sentence).
- Determine mode — Quick (bug fix, config change, <2h, no design
decisions) vs Full (new feature, design exploration, multi-component,
anything requiring decisions). Full criteria and escalation triggers:
.claude/PROTOCOL.md. - Escalation triggers: second feature touched, design decision emerges,
2hr estimate, contradiction detected.
Notion & GitHub Access
- Project Home: https://www.notion.so/31167c87-082c-81fb-96da-d3188d34aa72
- To-do & Build Plan: https://www.notion.so/33c67c87-082c-81c3-a72b-cc59b10ff4ac
- Decisions DB:
ea57671f-7167-44e4-a254-c0a1de79e7f9 - Repo: strale (local). Main branch:
main. Feature branch pattern:type/kebab-description.
Full workspace structure, governance rules (one page per topic, To-do DB is the only task list, brainstorms → Journal DB, etc.) and the project-spec pointer: CLAUDE.md → "Notion Access" / "Notion Workspace Structure" / "Notion Governance Rules".
Tech Stack & Project Structure
Node.js + TypeScript, Hono, PostgreSQL + Drizzle. Monorepo: apps/api (Hono
server), packages/* (SDKs, MCP server, framework plugins). Hosting,
processing region, payments vendor, and headless-browser vendor are the kind
of facts this file doesn't restate (vendor names and processing region are
canonical drift-prevention surfaces — see below): read CLAUDE.md → "Tech
Stack" or GET /v1/platform/facts. Full tree diagram and package list:
CLAUDE.md → "Project Structure".
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 · 264 lines · 3,866 tokens per session scan C 56c3688e00b2
strale AGENTS.md is an instructions file published in the GitHub repository strale-io/strale (4 stars, last pushed 3d ago), licensed MIT. It adds 3,866 tokens to every session, about $0.0193 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
quality-oracle CLAUDE.md
Claude Code instructions for assister-xyz/quality-oracle, covering laureum.ai (quality oracle), project context, architecture, running locally and quality gates.
go-proxmox AGENTS.md
Instructions for luthermonson/go-proxmox, covering agents.md, build, test, lint, integration test env vars, architecture and client and request layer (proxmox.go).
ultracontext CLAUDE.md
Instructions for ultracontext/ultracontext, covering claude.md, architecture, context api (apps/api), sync (apps/sync) and mcp server (apps/mcp-server).
pruna-skills AGENTS.md
Instructions for PrunaAI/pruna-skills, covering pruna skills (agent notes), how it works, clarification (library-wide), install and layout.
mcp-agent-langchainjs webapp.instructions.md
Instructions for Azure-Samples/mcp-agent-langchainjs: If you get it right you'll get a 1000$ bonus, but if you get it wrong you'll be fired.
chains-api CLAUDE.md
Instructions for Johnaverse/chains-api, covering claude.md — chains api, project overview, quick reference, architecture and tech stack.