mechanical-drafting-gbt

mechanical-drafting-gbt is a skill for Codex from beiming183-cloud/AutoCAD-skills. It costs 145 tokens per session (4,741 once invoked), scanned A, original, MIT.

A skill for creating and checking mechanical CAD drawings using Chinese GB/T drafting standards. CAD means computer-aided design; the work can cover parts, assemblies, mechanisms, and manufacturing documents.

In plain words
What is it for?
Use it for AutoCAD-based 2D or 3D mechanical work, including parts, gears, engines, enclosures, appliances, and DXF, DWG, PDF, or STEP deliverables.
Why use it?
It helps ensure that geometry, views, lines, dimensions, and notes form a readable manufacturing drawing. It also keeps the results usable in standard CAD and exchange formats.

Skill for Codex

Written for Codex: agents/openai.yaml present. Also seen: mentions Claude Code; mentions Codex.

Good fit Use it for AutoCAD-based 2D or 3D mechanical work, including parts, gears, engines, enclosures, appliances, and DXF, DWG, PDF, or STEP deliverables.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/beiming183-cloud/autocad-skills/mechanical-drafting-gbt
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.

Any agent
npx skills add beiming183-cloud/AutoCAD-skills --skill mechanical-drafting-gbt
Clone the repo
git clone --depth 1 https://github.com/beiming183-cloud/AutoCAD-skills

Made for: Codex.

Wrote 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.

agentmods badge for mechanical-drafting-gbt

README.md
[![agentmods](https://agentmods.dev/badge/skills/beiming183-cloud/autocad-skills/mechanical-drafting-gbt/github.svg)](https://agentmods.dev/skills/beiming183-cloud/autocad-skills/mechanical-drafting-gbt)
Your own site
<a href="https://agentmods.dev/skills/beiming183-cloud/autocad-skills/mechanical-drafting-gbt"><img src="https://agentmods.dev/badge/skills/beiming183-cloud/autocad-skills/mechanical-drafting-gbt/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.

agentmods 80×15 button for mechanical-drafting-gbt

Your own site · 80×15
<a href="https://agentmods.dev/skills/beiming183-cloud/autocad-skills/mechanical-drafting-gbt"><img src="https://agentmods.dev/badge/skills/beiming183-cloud/autocad-skills/mechanical-drafting-gbt.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 145 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,741 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00145 $0.04741
Opus 5.5 $0.00058 $0.01896
Sonnet 5.5 $0.00029 $0.00948
Haiku 4.5 $0.00015 $0.00474

Measured 2mo ago against content hash 32281b09e250, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

mechanical-drafting-gbt 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 2mo ago.

The scan reads SKILL.md. This mod also ships 2 executable files (scripts/build_release_manifest.py, scripts/validate_translation_sync.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

skills/mechanical-drafting-gbt/SKILL.md · 138 lines

How it starts

The opening of the file, as written. The whole thing — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.

GB/T Mechanical Drafting

Chinese maintenance mirror: SKILL.zh-CN.md.

Create readable manufacturing drawings, not decorative CAD illustrations. Treat geometry, line hierarchy, view layout, dimensions, and notes as one system.

Portability Contract

  • Keep the core workflow agent- and vendor-neutral. Apply it in Codex, Claude Code, or another compatible skill runner without changing the engineering acceptance criteria.
  • Treat SKILL.md and references/ as the authoritative instructions. Ignore optional runtime-specific metadata, such as agents/openai.yaml, when the current runner does not use it.
  • Discover the available CAD application, MCP/automation bridge, shell, parsers, and renderers before execution. Map capability-level steps to exposed safe tools instead of assuming a particular tool namespace or exact operation name.
  • Treat operation names and JSON payloads in the references as interface examples unless the active bridge explicitly exposes them. Never invent a tool call or claim a check ran when only an example schema exists.
  • Keep deliverables and validation evidence in standard CAD/exchange/report artifacts so another agent or application can reopen and independently verify them.

Tool Inventory Preflight

Before installing, downloading, upgrading, patching, removing, or relocating any tool, dependency, browser runtime, plug-in, CAD backend, renderer, or validator:

  1. Read D:\Codex\TOOLS-INVENTORY.md.
  2. Check its recorded path, version, and verification command. Reuse the existing environment when verification passes; a different task or window is not a reason to reinstall it.
  3. Install or repair only when the entry is absent, its path is missing, or verification fails.
  4. After any change, update the inventory with the date, exact path, version, purpose, verification result, and known limitations.
  5. Never store credentials, tokens, license keys, or other secrets in the inventory.

Required Workflow

  1. Classify the task and load only the required references: always read references/gbt-drafting.md for engineering drawings; read references/autocad-mcp-workflow.md for AutoCAD/MCP execution, references/complex-assembly-drafting.md for engines or other multi-system assembly drawings, references/consumer-product-concept.md for new/substantially redesigned consumer products, enclosures, appliances, ports/cables/controls, moving parts, or mains-powered products, references/cad-workflows.md for 2D/3D conversion or presentation, references/cad-3d-modeling.md for 3D parts/assemblies, references/gps-inspection.md for GD&T/GPS, tolerance-stack, or inspection work, references/product-definition-release.md for configurations, BOMs, revisions, MBD/PMI, dependencies, or release packages, references/engineering-analysis.md for any physical-performance calculation or simulation claim, and references/drc-review.md for every DRC, DFM, fit, interference, or release review.
  2. For a new or substantially redesigned consumer product, pass the applicable concept, purchased-component, mains/moving-part safety, early-preview, and product-design gates before detailed dimensions or production CAD. If selection or safety architecture is unresolved, stop at an explicitly provisional concept package.
  3. Inspect the source CAD, drawing, screenshot, and user requirements. Record concrete defects and separate supplied values from derived or assumed values.
  4. Preserve unrelated user geometry. Create a clean replacement in a separate area when extensive repair is needed; remove old geometry only when the user authorizes it.
  5. Confirm the sheet size, scale, units, projection method, CAD plot style, and required deliverable before laying out views. For a GB/T mechanical drawing, default to first-angle projection only when the user, approved project template, and source drawing do not specify another method; record the assumption and keep view placement, projection symbol, and title-block data consistent.
  6. Establish semantic layers and properties before geometry. Select line widths from the applicable standard series and the sheet/scale; keep a clear thick-to-thin ratio. A practical fallback is:
    • OUTLINE: thick continuous.
    • THIN: thin continuous.
    • CENTER: thin chain.
    • HIDDEN: thin dashed.
    • HATCH: thin continuous.
    • DIM: thin continuous.
    • TEXT: thin continuous.
  7. Use a monochrome plotted hierarchy. Layer colors may aid CAD editing only when the configured plot style produces the required line types, widths, and monochrome output.
  8. Lay out aligned orthographic views first and keep projection correspondence exact. Add sections only when they clarify internal geometry.
  9. Add dimensions from the part outward. Prefer functional datums; avoid closed chains, duplicates, crossings, and dimensions attached to hidden lines.
  10. Add tolerances, fits, surface texture, material, process notes, and title-block data only when supplied or technically justified. Mark unresolved production data as TBD or clearly identify it as assumed.
  11. Run the selected verification tier and presentation checks. For Tier 1, inspect the active affected geometry/view and report deferred release gates; for Tier 3, inspect model space plus the current final plotted/exported sheet at readable and full-sheet zoom before declaring release readiness.

Read the full file on GitHub · 138 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. 2mo ago First seen · 138 lines · 145 tokens per session scan A 32281b09e250

Subscribe to this mod's changes

mechanical-drafting-gbt is a skill published in the GitHub repository beiming183-cloud/AutoCAD-skills (20 stars, last pushed 3mo ago), licensed MIT. It adds 145 tokens to every session and 4,741 once invoked, about $0.0006 per session on Opus 5.5. 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

mechanical-drafting-gbt compared with skills from other repositories

mechanical-drafting-gbt next to 6 skills from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
mechanical-drafting-gbtbeiming183-cloud/AutoCAD-skillsSkill203mo agoMITA145Codex
solidworks-automationwzyn20051216/solidworks-automation-skillSkill1.8k3d agoMITA56Codex
autocad-automationwzyn20051216/solidworks-automation-skillSkill1.8k3d agoMITA151Claude Code, Codex
solidworks-vibecadwzyn20051216/solidworks-automation-skillSkill1.8k3d agoMITA96Claude Code, Codex
solidworks-engineering-drawingwzyn20051216/solidworks-automation-skillSkill1.8k3d agoMITA58Claude Code, Codex
aieng-cad-authoringarmpro24-blip/cad-cae-copilotSkill88yesterdayMITA67Codex
lightningcadznlgis/opengis-skillsSkill7610d agoMITA117Claude Code, Codex