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/ayeshlk/archsmith/agents-mdgit clone --depth 1 https://github.com/ayeshLK/archsmithWhat 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 | $0.09374 | $0.09374 |
| Opus 5 | $0.04687 | $0.04687 |
| Sonnet 5 | $0.01875 | $0.01875 |
| Haiku 4.5 | $0.00937 | $0.00937 |
Grade A, and why
archsmith 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 — 183 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI coding agents (Claude Code, Cursor, Copilot, Codex, etc.) working on the ArchSmith repository itself.
Scope: this file configures agents contributing to this repo. It complements CONTRIBUTING.md (the human-facing build/governance doc) rather than replacing it — read both, and don't duplicate one into the other if you're updating them.
What this is
A renderer, CLI, and MCP server that turn a validated JSON IR (intermediate representation) into an SVG diagram, in one fixed layered/swimlane house style. See README.md for the full pitch and the IR shape. The short version: @archsmith/renderer is a pure function (IR in, SVG out, no LLM call inside it) — interpretation (sketch/description → IR) is deliberately someone else's job, not this repo's.
Architecture at a glance
npm workspaces monorepo. Four packages, strict one-way dependency chain — each package only knows about the one directly below it:
@archsmith/cli ─┐
├──▶ @archsmith/renderer ──▶ @archsmith/schema
@archsmith/mcp-server ─┘
packages/schema— the JSON Schema (diagram-schema.json) plus governed registries (registries/{colors,icons,sub-layers}.json). Registry entries are house-style contracts, not per-diagram decisions.packages/renderer— pure function: validated IR in, SVG string out. No I/O beyond reading its own bundled font. No LLM call anywhere inside it. Layout is organized aslayout/(per-column assembly),boxes/(individual shapes),text/(font measurement + wrapping via the embedded Arimo font, usingfontkit), andsvg/(node model + serialization + font embedding). TypeScript project references (tsc -b) handle inter-package build order — no manual ordering needed.packages/cli— thearchsmithbinary. Thincommander-based wrapper aroundrender()/validate().renderexits 1 on validation errors, 2 on rendering errors. Also has theauthorsubcommand, an interactive Ink-based wizard — seearchsmith author, below.packages/mcp-server— thearchsmith-mcpbinary. A sibling of the CLI, not a wrapper: callsrender()/validate()from@archsmith/rendererdirectly.server.tsbuilds theMcpServerwhileindex.tsis a thin entrypoint that wires it toStdioServerTransport; the split exists soserver.test.tscan connect a real MCPClientover an in-memory transport pair. Never write to stdout — stdout is the JSON-RPC channel and a strayconsole.log(even for debugging) corrupts the protocol stream. Useconsole.errorfor diagnostics.
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 · 183 lines · 9,374 tokens per session scan A 5c4566ec9b00
archsmith AGENTS.md is an instructions file published in the GitHub repository ayeshLK/archsmith (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 9,374 tokens to every session, about $0.0469 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
joint CLAUDE.md
Instructions for clientIO/joint, covering claude.md, project overview, common commands, install dependencies and build distribution files (runs grunt).
thesvg AGENTS.md
AGENTS.md instructions for glincker/thesvg, covering thesvg - ai review guidelines, svg icon submission rules, required, fill and color rules and icons.json registry.
neodx AGENTS.md
Instructions for secundant/neodx, covering neodx, session rules, package layers, command vocabulary (current) and verification.
drawio-ai-kit AGENTS.md
Instructions for sparklabx/drawio-ai-kit, covering repository guidelines, tech stack & runtime and code conventions & common patterns.
pr-lens AGENTS.md
Instructions for coldteadotai/pr-lens, covering pr lens — agent instructions, hard requirements, general, 1. think before coding and 2. simplicity first.
inkscape-mcp AGENTS.md
Instructions for sandraschi/inkscape-mcp, covering inkscape-mcp — agent guide, entry points, standalone desktop installer required, standards and key files.