agentsys CLAUDE.md

A set of project instructions for agent-sh and agentsys, a modular runtime for AI coding agents. It defines priorities, model choices, code-quality rules, testing expectations, and development commands.

In plain words
What is it for?
Use it when agent-sh or agentsys is developing or changing code, especially when deciding which model to use, running tests, or preparing a pull request.
Why use it?
It keeps agents consistent about quality, safety, model selection, and verification while they work on the project.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/agent-sh/agentsys/claude-md
Clone the repo
git clone --depth 1 https://github.com/agent-sh/agentsys
Per session 360 This file is loaded in full into every session.
When invoked 360 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.00360 $0.00360
Opus 5 $0.00180 $0.00180
Sonnet 5 $0.00072 $0.00072
Haiku 4.5 $0.00036 $0.00036

Measured 3d ago against content hash a15b7fc928e3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agentsys CLAUDE.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.

Origin

Copies of this mod

2 near-identical copies found in the catalogue:

CLAUDE.md · 43 lines

What it actually says

agentsys

A modular runtime and orchestration system for AI agents - works with Claude Code, OpenCode, and Codex CLI

Critical Rules

  1. Plain text output - No emojis, no ASCII art. Use [OK], [ERROR], [WARN], [CRITICAL] for status markers.
  2. No unnecessary files - Don't create summary files, plan files, audit files, or temp docs.
  3. Task is not done until tests pass - Every feature/fix must have quality tests.
  4. Create PRs for non-trivial changes - No direct pushes to main.
  5. Always run git hooks - Never bypass pre-commit or pre-push hooks.
  6. Use single dash for em-dashes - In prose, use - (single dash with spaces), never --.
  7. Report script failures before manual fallback - Never silently bypass broken tooling.
  8. Token efficiency - Save tokens over decorations.

Model Selection

Model When to Use
Opus Complex reasoning, analysis, planning
Sonnet Validation, pattern matching, most agents
Haiku Mechanical execution, no judgment needed

Core Priorities

  1. User DX (plugin users first)
  2. Worry-free automation
  3. Token efficiency
  4. Quality output
  5. Simplicity

Dev Commands

npm test          # Run tests
npm run validate  # All validators

References

Changes

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.

  1. 3d ago First seen · 43 lines · 360 tokens per session scan A a15b7fc928e3

Subscribe to this mod's changes

agentsys CLAUDE.md is an instructions file published in the GitHub repository agent-sh/agentsys (980 stars, last pushed 6d ago), licensed MIT. It adds 360 tokens to every session, about $0.0018 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.