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/jdilla1277/agentcad/claude-mdgit clone --depth 1 https://github.com/jdilla1277/agentcadWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/jdilla1277/agentcad/claude-md)<a href="https://agentmods.dev/instructions/jdilla1277/agentcad/claude-md"><img src="https://agentmods.dev/badge/instructions/jdilla1277/agentcad/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.01232 | $0.01232 |
| Opus 5 | $0.00616 | $0.00616 |
| Sonnet 5 | $0.00246 | $0.00246 |
| Haiku 4.5 | $0.00123 | $0.00123 |
Grade A, and why
agentcad 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 6d 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 — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
agentcad
agentcad is a local CLI and MCP server for AI-agent CAD workflows.
Repo Identity
This checkout is the public jdilla1277/agentcad repo. It is for
externally safe package code, public docs, and examples.
Do not open PRs here for internal planning material: PRDs, roadmap notes, marketing drafts, promotional plans, feedback logs, launch notes, private operational context, or anything that should not be public.
If the user asks for internal planning, promotion work, or website work for
agentcad.dev, use the internal repo (jdilla1277/agentcad-internal) instead.
Before creating any PR, run git remote -v and confirm whether the target repo
is public or internal.
To get a working tree for the internal repo without disturbing this checkout:
git -C ~/conductor/repos/agentcad-internal worktree add <path> -b <branch> origin/main
The website lives at agentcad-site/ in that repo.
Development
Use Python 3.10-3.12. CadQuery/OpenCascade does not support Python 3.13+.
python3.12 -m venv .venv
source .venv/bin/activate
pip install -e ".[mcp,dev]"
pytest
Product Contract
- Commands return structured JSON.
agentcad runcreates versioned output directories and records metadata.agentcad docsandagentcad --helpare part of the agent-facing API.- Keep error messages concise and actionable for coding agents.
- Prefer local, deterministic workflows over hosted dependencies.
- The command list and the
agentcad docssection list are pinned bytests/test_public_surface.py. Adding or removing either fails that test until you regenerate the snapshot withAGENTCAD_UPDATE_SURFACE=1 pytest tests/test_public_surface.pyand update the hand-written mirror at agentcad.dev/docs. The website is in a different repo, so nothing else will catch that drift.
Validation
For changes to agent-facing contracts, CLI workflows, docs/help output, viewer behavior, or app-like user flows, run a narrow sub-agent friction check before opening or finalizing the PR. Ask the sub-agent to behave like a fresh agent: read the docs, use the feature end-to-end in a scratch project, and report confusing behavior or mismatches between docs and reality.
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.
- 6d ago First seen · 117 lines · 1,232 tokens per session scan A 9d99e15ead4e
agentcad CLAUDE.md is an instructions file published in the GitHub repository jdilla1277/agentcad (111 stars, last pushed yesterday), licensed Apache-2.0. It adds 1,232 tokens to every session, about $0.0062 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
partcad AGENTS.md
AGENTS.md instructions for partcad/partcad, covering partcad, the packages, all inside the one wheel, everything else, development process and where commands run.
build123d-mcp CLAUDE.md
Claude Code instructions for pzfreo/build123d-mcp, covering build123d-mcp, running tests, running the server, project structure and adding a new tool.
cad-cae-copilot copilot-instructions.md
Copilot instructions for armpro24-blip/cad-cae-copilot, covering github copilot — aieng workspace and essentials.
cad-cae-copilot CLAUDE.md
Claude Code instructions for armpro24-blip/cad-cae-copilot, covering claude code — aieng workspace and tl;dr.
CADCLAW AGENTS.md
AGENTS.md instructions for sunnyday-technologies/CADCLAW, covering agents.md — guidance for ai assistants using cadclaw, what cadclaw is, and what it is not, the core rule: place authored parts, don't generate them, why this rule exists and what you may generate.
IntentForge AGENTS.md
Instructions for jzjzzzzzzz/IntentForge, covering agent instructions, current phase, engineering rules and cad rules.