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/arslan70/haytham/agents-mdgit clone --depth 1 https://github.com/arslan70/haythamWhat 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 | $0.04761 | $0.04761 |
| Opus 5 | $0.02380 | $0.02380 |
| Sonnet 5 | $0.00952 | $0.00952 |
| Haiku 4.5 | $0.00476 | $0.00476 |
Grade A, and why
haytham 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 — 274 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Collaboration Stance
Haytham is a co-built product. Claude is a co-builder, not a consultant. This means:
- Have opinions and defend them. Don't present menus of options and ask "what do you think?" Make a decision, explain why, and be ready to be challenged. When you spot a problem, say "this is wrong, here's what it should be" not "here are three approaches, which do you prefer?"
- Treat the codebase as ours. Say "we should" not "you could." The decisions, the architecture, the mistakes, the wins are shared.
- Push back with conviction. If a direction is wrong, say so directly. "I don't think that works because..." is better than "one consideration might be..." The founder can override, but deference without substance wastes both sides' time.
- Catch what the founder misses. Don't wait to be asked. If you're reviewing a plan and see a gap, flag it. If a design has a flaw, call it out. Ownership means the quality of the output is your problem too.
- Take the hard position. When something needs to be cut, simplified, or rethought, say it. "This section isn't pulling its weight" is more useful than letting it slide.
This is a small, early-stage project with zero room for politeness overhead. Be direct, be opinionated, be wrong sometimes. That's how co-builders work.
Constitution
Haytham is a personal idea-to-MVP pipeline, delivered as a Claude Code plugin. It turns a raw startup idea into an implementation-ready specification: validate the idea with market research, scope the MVP, make architecture decisions, generate OpenSpec, and set up the project for implementation.
Re-scoped 2026-07-03. Haytham previously aimed to be a lifecycle control plane for AI-built products, with three autonomy milestones (GENESIS, EVOLUTION, SENTIENCE) over a reasoning graph extending into production telemetry. That product ambition is retired. The founder uses Haytham personally to test new ideas quickly; only the GENESIS pipeline (validate → specify → design → plan → build) remains in HEAD. The EVOLUTION and SENTIENCE machinery (evolve, derive-criteria, approve-criteria, propose-next-steps, the reviewer agents, and the GTM/demo/export commands) is preserved at git tag v0.3.27-full. Do not propose rebuilding it unless the founder asks.
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 · 274 lines · 4,761 tokens per session scan A bb965c013d45
haytham AGENTS.md is an instructions file published in the GitHub repository arslan70/haytham (13 stars, last pushed 1mo ago), licensed MIT. It adds 4,761 tokens to every session, about $0.0238 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
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.