pi-arcweld AGENTS.md

A project instruction document for pi-arcweld, which contains local rules for a Pi workspace built around the pi-mono source code.

In plain words
What is it for?
Use it when developing Pi extensions, MCP servers, or other components in the pi-arcweld workspace.
Why use it?
It tells coding agents where extensions and MCP servers belong, how they are deployed, and which repository changes are allowed.

Instructions file for CodexOpenCode

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/appautomaton/pi-arcweld/agents-md
Clone the repo
git clone --depth 1 https://github.com/appautomaton/pi-arcweld

Made for: Codex, OpenCode.

Per session 1,119 This file is loaded in full into every session.
When invoked 1,119 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 $0.01119 $0.01119
Opus 5 $0.00560 $0.00560
Sonnet 5 $0.00224 $0.00224
Haiku 4.5 $0.00112 $0.00112

Measured 2d ago against content hash c25e6296503e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pi-arcweld 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.

AGENTS.md · 47 lines

How it starts

The opening of the file, as written. The whole thing — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Local Pi Workspace Rules

This repository wraps the upstream pi-mono source as a pinned Git submodule.

Workspace Layout

  • pi-mono/ is an upstream Git submodule, pinned by the root repository. Keep it free of local build/dependency state and do not add it to a root npm workspace.
  • extensions/ is the single repository root for locally maintained Pi extensions. Keep every curated extension there. Global extensions may be symlinked into ~/.pi/agent/extensions/ or registered as local-path packages in ~/.pi/agent/settings.json; document the chosen loading method in extensions/README.md.
  • mcp-servers/ holds the canonical source for locally maintained MCP servers. They are standalone processes, not Pi extensions, and each owns its package and release workflow. Develop and test them here, then deploy tested versioned releases under ~/.local/mcps/<server>/releases/; Pi must run the ~/.local/mcps/<server>/current/ launcher registered by absolute path in ~/.pi/agent/mcp.json, not the mutable Git checkout. See mcp-servers/README.md.
  • Chrome DevTools MCP is external to this repository. Its single canonical root is ~/.local/mcps/chrome-devtools/; read its AGENTS.md and README.md before changing the integration. The Chrome browser remains under ~/.local/browsers/chrome/.
  • system-instruction/ contains the canonical APPEND_SYSTEM.md, symlinked from ~/.pi/agent/APPEND_SYSTEM.md. Machine-specific reference captures under baseline/ are local and ignored. Do not create a project-level .pi/APPEND_SYSTEM.md, because it would replace the global append file for this project rather than layer with it.
  • scripts/ contains local runtime build, validation, and upstream-update scripts.
  • docs/ contains the static GitHub Pages site; keep its claims and paths consistent with the repository documentation.
  • references/ contains local independent repositories retained for comparison or tooling. It is ignored and is not part of the public repository.
  • build/pi-agent/runtime/ is the runnable local Pi artifact.
  • build/pi-agent/work/ is temporary build state and is deleted by default after a successful build.
  • build/pi-agent/artifacts/ contains local package tarballs used to assemble the runtime.
  • archive/ is for local logs, session exports, and other files that should not clutter the workspace root.

Read the full file on GitHub · 47 lines

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. 2d ago First seen · 47 lines · 1,119 tokens per session scan A c25e6296503e

Subscribe to this mod's changes

pi-arcweld AGENTS.md is an instructions file published in the GitHub repository appautomaton/pi-arcweld (6 stars, last pushed 3d ago), licensed MIT. It adds 1,119 tokens to every session, about $0.0056 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.

Related

Other instructions, from other repositories