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/magolor/blueprint/agents-mdgit clone --depth 1 https://github.com/Magolor/BlueprintWhat 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 | $0.00583 | $0.00583 |
| Opus 5 | $0.00292 | $0.00292 |
| Sonnet 5 | $0.00117 | $0.00117 |
| Haiku 4.5 | $0.00058 | $0.00058 |
Grade A, and why
Blueprint 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 — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Blueprint Agent Guide
Purpose and authority
The typescript branch is the active Blueprint product line. It is a strict starter for a small Node SDK and CLI package.
Use this authority order:
- Current user instructions and this guide.
- Shipped code, tests, package artifacts, and CI.
README.en.mdanddocs/README.md.- Current plans and reports.
package.json owns direct dependencies and commands. pnpm-lock.yaml owns resolved dependency state. Use pnpm only.
Work loop
- Run
rtk pnpm tasks --readyfor non-trivial work. - Claim an existing task before you create one.
- Keep resumable state only in
docs/tasks.yaml. - Use one linked plan for architectural or multi-slice work.
- Preserve unrelated changes and other worktrees.
- Run focused checks during implementation.
- Run
rtk pnpm checkbefore closeout. - Update canonical docs and
docs/DEVLOG.md. - Remove completed or canceled queue rows.
The queue supports draft, ready, active, blocked, and postponed. Record cancellation during closeout, then remove the row.
TypeScript architecture
- Keep one cohesive native ESM package until a real boundary earns another package.
- Use
src/index.tsas the only public SDK boundary. - Use
src/cli.tsas a thin adapter over the SDK. - Validate external values from
unknown. - Use direct functions for stateless behavior. Do not invent a template-domain object.
- Keep strict compiler options enabled.
- Add no framework, GUI, service, database, plugin, or workspace without a concrete requirement.
Commands
Prefix local agent commands with rtk.
rtk pnpm install --frozen-lockfile
rtk pnpm check:fast
rtk pnpm check
rtk pnpm build
rtk pnpm package:check
pnpm check is the complete code, documentation, and package gate.
Branch and skill contract
typescript is active and is the hosted default. python is the compatibility line. These are the only long-lived branches.
Product files can differ. The .agents/skills/heaven-style/ tree must be byte-identical on every configured product branch. The default branch set is python typescript.
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 · 69 lines · 583 tokens per session scan A 70c95cad98da
Blueprint AGENTS.md is an instructions file published in the GitHub repository Magolor/Blueprint (5 stars, last pushed 6d ago), licensed MIT. It adds 583 tokens to every session, about $0.0029 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
py-launch-blueprint AGENTS.md
Instructions for smorinlabs/py-launch-blueprint, covering agents.md, required tools, canonical commands, verification flow before commit/pr and pull request review comments — merge guard (hard, enforced).
py-launch-blueprint CLAUDE.md
Instructions for smorinlabs/py-launch-blueprint: Canonical agent guidance lives in AGENTS.md (single source of truth — edit that file, not this one) and is imported below. Add ONLY Claude-specific notes to this file.
turbo-start-sanity CLAUDE.md
Instructions for robotostudio/turbo-start-sanity, covering claude.md, project overview, commands, development (both apps) and individual apps.
vercel-action CLAUDE.md
Instructions for amondnet/vercel-action, covering claude.md, project overview, package manager, essential commands and development.
cookiecutter-mlops-package AGENTS.md
Instructions for fmind/cookiecutter-mlops-package, covering agents.md, project overview, setup & core commands, definition of done and conventions & idioms.
seedkit CLAUDE.md
Instructions for viewflow/seedkit, covering seedkit, writing reference files, changelog, maintaining testcases and submodule workflow.