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 skills add armpro24-blip/cad-cae-copilot --skill aieng-closed-loop-copilotgit clone --depth 1 https://github.com/armpro24-blip/cad-cae-copilotWrote 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/skills/armpro24-blip/cad-cae-copilot/aieng-closed-loop-copilot)<a href="https://agentmods.dev/skills/armpro24-blip/cad-cae-copilot/aieng-closed-loop-copilot"><img src="https://agentmods.dev/badge/skills/armpro24-blip/cad-cae-copilot/aieng-closed-loop-copilot/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/armpro24-blip/cad-cae-copilot/aieng-closed-loop-copilot"><img src="https://agentmods.dev/badge/skills/armpro24-blip/cad-cae-copilot/aieng-closed-loop-copilot.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00054 | $0.00721 |
| Opus 5 | $0.00027 | $0.00360 |
| Sonnet 5 | $0.00011 | $0.00144 |
| Haiku 4.5 | $0.00005 | $0.00072 |
Grade A, and why
aieng-closed-loop-copilot 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 10d 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 — 46 lines — stays where its author put it; the contents beside it link to each section on GitHub.
aieng-closed-loop-copilot
MCP-first closed-loop discipline for recommend → edit → re-simulate → compare. Use this only when a baseline project has CAD, CAE setup/results, and explicit design targets or user-stated metrics.
Purpose
- Bound design iteration before changing geometry.
- Prefer safe parameter edits over full regeneration when editable parameters exist.
- Re-run CAE before claiming improvement.
- Stop on target met, no evidence, approval denial, or budget exhaustion.
- For repeatable loops, follow the relevant entry in
../engineering_skill_contracts.json:cad-mod-propose-verify,solver-run-orchestrate,design-target-review, orevidence-report-synthesize.
Required workflow
- Define the target, metric, and iteration budget (default: 3).
- Inspect baseline with
aieng.agent_context,aieng.inspect_package, and result/design-target summaries when present. - Discover editable dimensions with
cad.list_editable_parametersand inspect local/global scope. - Choose one change per iteration. Prefer
cad.edit_parameter; usecad.execute_build123d,cad.replace_part, orcad.remove_partonly when geometry must change structurally. - After CAD mutation, inspect
regression_diff(topology drift) andcritique_diff(afail/warnverdict means the edit worsened manufacturability) and refresh semantics if needed. - Run
cad.critiquefor engineering parts before CAE claims. - Prepare CAE with
cae.prepare_solver_run, generate deck withcae.generate_solver_inputif needed, then callcae.run_solveronly through approval. - Extract metrics with
cae.extract_solver_results, optionallycae.extract_field_regions, and refresh summaries before comparing.
Hard rules
- Respect
[APPROVAL REQUIRED]tools; ifAIENG_MCP_BLOCK_APPROVAL_TOOLS=1blocks mutation/solver execution, report the block and stop. - Do not chain multiple CAD changes before re-simulation unless the user explicitly accepts the loss of attribution.
- Do not claim improvement until post-change solver/result evidence exists.
- Read the
credibilitytier on every result:executed_solver_resultoutranksproxy_assembly_resultoutrankssurrogate_predictionoutrankscritique_finding. Compare like-for-like and never present a lower tier as solver-verified. - Surrogate proposals are advisory, not evidence: never report a predicted number without its envelope — each prediction carries
uncertainty_std/predicted_score_band, and the proposal set carries a leave-one-outvalidationerror band. They guide search only and never substitute for a solver rerun. - A critique pass or parameter range check is not production certification.
- Stop if target is already met, no safe editable variable exists, no solver evidence can be produced, approval is denied, or budget is exhausted.
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.
- 10d ago First seen · 46 lines · 54 tokens per session scan A 28efa072484c
aieng-closed-loop-copilot is a skill published in the GitHub repository armpro24-blip/cad-cae-copilot (58 stars, last pushed 2d ago), licensed MIT. It adds 54 tokens to every session and 721 once invoked, about $0.0003 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 skills, from other repositories
kernelcad-trace-from-image
Convert pixel-space features from a reference photo into normalized [0,1] waypoints, then into mm coordinates feedable to path().spline / path().nurbsSegment. Uses the tracefromimage MCP tool. Load when your blockout is in place and the next step is an organic-curve outline (eyewear brow, ergonomic handle, sneaker…
simplecadapi
Build, assemble, inspect, reconstruct, and export parametric CAD models with the SimpleCADAPI Python SDK. Use for SimpleCAD geometry modeling, constrained sketches, parts and assemblies, standard gears and bearings, STEP/BREP inspection and reconstruction, durable product packages, model JSON replay, and CAD backend…
b123d-modeling
Use this skill when asked to model, build, or modify a 3D part or assembly with build123d — from a text description, a technical drawing (image or PDF), dimensions in a spec, or an existing STEP/STL file.
kernelcad-authoring
Author or modify kernelCAD models in TypeScript. Scripts live in .kcad.ts files; the kernelCAD CLI (kernelcad evaluate and kernelcad export stl|step|dxf|3mf|glb -o ) executes them via an OpenCASCADE WASM kernel.
kernelcad-mcp
Load when kernelcad mcp is available and you need to inspect or edit a live model without re-running the full CLI on every change. The MCP server exposes dynamic introspection tools that operate on evaluated in-memory state.
kernelcad-nurbs
NURBS surfaces (nurbsSurface, surfaceFromCurves, surfaceFromBoundary, .thicken, .toShape) AND NURBS curves (nurbsCurve, spline3d, hermiteG2) AND multi-section sweeps (variableSweep) AND G1/G2 fillet continuity AND 2D NURBS path segments (path().spline / .nurbsSegment / .hermiteG2). Use for freeform geometry that…