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 JMMonte/agentic-digital-twin --skill manufacturability-loopgit clone --depth 1 https://github.com/JMMonte/agentic-digital-twinWrote 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/jmmonte/agentic-digital-twin/manufacturability-loop)<a href="https://agentmods.dev/skills/jmmonte/agentic-digital-twin/manufacturability-loop"><img src="https://agentmods.dev/badge/skills/jmmonte/agentic-digital-twin/manufacturability-loop/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/jmmonte/agentic-digital-twin/manufacturability-loop"><img src="https://agentmods.dev/badge/skills/jmmonte/agentic-digital-twin/manufacturability-loop.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00089 | $0.00850 |
| Opus 5 | $0.00044 | $0.00425 |
| Sonnet 5 | $0.00018 | $0.00170 |
| Haiku 4.5 | $0.00009 | $0.00085 |
Grade A, and why
manufacturability-loop 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 8d 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 — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The manufacturability / packaging loop
Manufacturability is a deliberate BLEND: judgment that only a human-or-agent can author, scored by automation that a human-or-agent cannot fake. Never let one silently replace the other — a wrong placement still parses as valid JSON, and a wrong boolean still passes watertight+volume>0. The solver is what proves the judgment actually fits.
The two layers
- AGENT layer —
mfg/judgment.json(you author this): how the design decomposes into FABRICATED parts vs COTS, the cable/fluid harness, the joints, and each item's PLACEMENT (position + envelope) and SOURCING. - AUTOMATION layer —
mfg/packaging_solve.py(this scores it): skin containment against the REAL inner ring, envelope COLLISIONS, MASS closure, CG, clearance to moving parts, and RETRIEVAL corridors. Exit 1 on FAIL.
The loop
-
Author judgment. Edit
mfg/judgment.json: list fabricated parts (process + material + est. mass), COTS (vendor PN + mass + pos + envelope + sourcing note), the harness, joints, and access openings. Use real datasheet envelopes — a too-small box that "looks placed" fails the density check. -
Score it.
uv run mfg/packaging_solve.py. Readmfg/packaging.md. -
Fix what's flagged. Each FAIL names the item and the violation:
- containment → the item pokes out the skin; move it inboard / down. Remember tumblehome: a side box can poke out the TOP corner too.
- collision → two envelopes intersect; re-place one.
- mass closure → the fabricated + COTS total is off MTOW; revisit the decomposition or masses.
- retrieval → a serviceable item isn't reachable through any opening, or a fixed box blocks the pull corridor; clear the channel.
-
Re-run. Repeat until 0 FAIL. WARN is allowed for honest, documented compromises; FAIL is not.
-
Verify the boolean, don't trust it. When the layout drives real geometry (cuts, mounts, decks), the build must export the as-built result and an audit must assert its expected bbox/volume/CG — because the kernel can silently no-op a boolean (see GUARDRAILS.md §1).
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.
- 8d ago First seen · 73 lines · 89 tokens per session scan A 9f229fe6ac1d
manufacturability-loop is a skill published in the GitHub repository JMMonte/agentic-digital-twin (2 stars, last pushed 2mo ago), licensed MIT. It adds 89 tokens to every session and 850 once invoked, about $0.0004 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 skills, from other repositories
cad-construction
Use after the design concept is settled (via design-first-iteration or equivalent) and the user is ready to translate it into parametric CAD geometry — component decomposition, parameter hierarchy, construction sequencing. Primary platform Fusion 360; the workflow also applies to SolidWorks, Onshape, FreeCAD…
bom-validation-workflow
Use when validating a Bill of Materials (BOM) for a Maker-3D project — an embedded controller project, a speaker/audio build, a smart-home build, or any other physical build with more than 5 components. Trigger on phrases like "check the BOM", "validate the parts list", "review the parts list", "missing components"…
victron-cerbo-modbus-device-onboarding
Use when a new Victron Modbus-TCP device (EV Charging Station, energy meter, inverter) is connected to the Cerbo GX LAN but does not appear in the Cerbo UI or D-Bus, despite being reachable via ping. Trigger on phrases like "EVCS not visible in Cerbo", "new Modbus device does not appear", "Cerbo does not detect…
mechanical-design-principles
Guides mechanical design decisions toward elegant, manufacturable, serviceable solutions with a strong bias toward tool-less assembly, monolithic parts, and form-follows-function. Provides the design rules (the WHAT and WHY — overhang, wall thickness, snap-fit sizing, tolerances, material matching). Load when the user…
hsb-ip-packetizer
Choose or explain HSB Sensor RX packetizer fields for HOLOLINKdef.svh. Do not use for full defs, validation, or runtime APB programming.
jetson-diagnostic
Read-only Jetson health snapshot for identity, memory, GPU, thermal, power, storage, services, and top processes.