by capskov
Road distress and pavement demolition annotation Skill for AI agents using AutoCAD: evidence-based scope, engineering quantities, hatches, callouts and legends.
# Add to your Claude Code skills
git clone https://github.com/capskov/cad-distress-annotationGuides for using ai agents skills like cad-distress-annotation.
See how cad-distress-annotation compares with popular alternatives.
cad-distress-annotation is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by capskov. Road distress and pavement demolition annotation Skill for AI agents using AutoCAD: evidence-based scope, engineering quantities, hatches, callouts and legends. It has 0 GitHub stars.
cad-distress-annotation's catalog security scan is still queued. You can run an instant dependency and prompt-injection check now with the "Scan for vulnerabilities" button above.
Clone the repository with "git clone https://github.com/capskov/cad-distress-annotation" and add it to your Claude Code skills directory (see the Installation section above). cad-distress-annotation ships a SKILL.md manifest, so compatible agents can discover and load it automatically.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh cad-distress-annotation against similar tools.
No comments yet. Be the first to share your thoughts!
Available after the catalog security scan passes.
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
The deep catalog scan for this skill is still queued. Run an instant dependency check now instead.
See comparison
Convert confirmed engineering scope into editable native CAD outcomes. Reuse the existing metre-road rules; distress severity, treatment depth and field observations cannot be inferred from demolition history. 中文规则. Read engineering examples when these cases apply and capability limits before native execution.
Confirm DWG, units, model/layout space, existing frame/legend, survey/design versions, quantity-table version and current revisions. Resume the authorized drawing without rebuilding its frame or replaying earlier batches. Keep analysis sources read-only; preserve an appropriate pre-edit backup. Explicit source-edit authorization governs whether to use the original or a working copy.
Inspect actual connected tools and schemas. A CAD year or file on disk does not prove an MCP connection. Identify the target process/document/backend and supported entity semantics. Use bounded diagnostics. After an uncertain write, query the original request and existing handles; never switch tools to duplicate it. Old 2022 access details do not establish a 2026 route.
Record boundary, hatch, text, leader, original and archived handles for every group. Resume from native state and latest approved changes; do not restore ranges cancelled by the user.
Use colors/hatches with one-to-one category legend entries. Linear items use dedicated measurement polylines, point items use counts, and neither uses area units.
For ordinary single-line text, yellow dot, associated LEADER/MTEXT and finite placement, load the installed sibling cad-callout-pair-layout/SKILL.md; public rules. This Skill owns scope, not callout geometry. Preserve explicitly requested multiline notes.
Summarize tiny fragments and inspect origin before treating them as clipping/numerical remnants. Prefer merging confirmed remnants into an adjacent same-category face, never across structures. If the user authorizes label-only hiding, hide text and associated leader together while retaining boundaries/hatches/quantities. A historical strictly-less-than-0.1-square-metre threshold is case-specific, not a minimum valid distress size or a threshold for other drawing units.
Lay out new/affected labels using their own remaining finite budget; old groups are fixed obstacles. Report unresolved positions after the budget. Existing objects alone do not prove collision-free appearance.
Read back new objects and associations before hiding old labels. Archive only the identified objects, including obsolete labels/leaders/duplicate legends, to an off/frozen/non-plot layer. Preserve restoration and do not turn off a shared layer such as layer 0.
Check affected closure, hatch path/native area, structure text, associated handles, duplicates/omissions and visibility. Separate area accuracy from path/position: equal area does not prove coincident geometry. Routine area comparison within 0.001 square metres is acceptable for this workflow; choose positional/width/length tolerances separately for their engineering effect, not from the area threshold.
Record save state/DBMOD, file digest and exact checked scope. Save and reopen are separate evidence; do not claim reopen when it did not happen. Report engineering differences, manual visual items and backend issues separately. Do not expand a requested partial handoff into unrequested whole-drawing acceptance. If asked to close CAD, save authorized writable targets, close references/targets normally and verify exit.
This public rules package includes no native CAD writer, automatic installer or engineering fixtures. Foundation provides supported operations and execution governance; other verified task-authorized routes can execute additional operations but do not automatically share its journals/fences. The full historical workflow is not certified using Foundation v1 alone. See installation and validation.
中文 · Install · Compatibility · Validation · Release v0.1.0-rc.1
Evidence-based road distress and demolition annotation rules for AI agents working with AutoCAD. The first public rules preview preserves existing professional decisions and revision practices. It guides boundaries, combined-structure quantities, category hatches, length/count labels and legends; it does not determine distress treatment without engineering evidence.
This is an installable Agent instruction package, not a CAD plugin, geometry engine or MCP server. No Python runtime is needed to read this Skill. Actual CAD execution needs verified native tools. Foundation supplies supported bounded operations via MCP/CLI; its current profile does not cover the complete demolition workflow. Other backends require explicit capability/identity checks and task authorization.
| Task | Components |
|---|---|
| Engineering scope review only | This Skill and the project's cleared evidence; no mandatory Foundation dependency. |
| This project's native callout presentation | This Skill + cad-callout-pair-layout + a verified suitable CAD backend. |
| Supported native operations with Foundation governance | Add cad-agent-foundation; verify missing operations before dependent writes. |
| Basic CAD operations unrelated to distress | Neither engineering Skill is mandatory. |
Metre-road rules and historical multi-tool AutoCAD 2026 evidence are documented separately from this publication. AutoCAD 2022 and complete Foundation-only execution are not certified. Read the synthetic decision examples and existing anonymous visual case; neither is a replay dataset or new native acceptance.
Download the release ZIP and follow manual installation. English SKILL.md and Chinese SKILL.zh-CN.md are included; keep resource folders together. Maintainer: capskov. Owned rules/documentation: Apache-2.0, see NOTICE.
Related terms: AutoCAD, CAD annotation, road distress, pavement demolition, civil engineering, engineering quantities, AI agent Skill. Independent project; no Autodesk endorsement.