Agentcad Skill
AdvisoryAudited by Static analysis on May 6, 2026.
Overview
No suspicious patterns detected.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Generated or edited CAD scripts will be run on the user's machine, so unsafe Python code could matter if the CLI does not sandbox execution.
The core workflow intentionally runs local Python/CadQuery scripts through the agentcad CLI. This is disclosed and purpose-aligned, but it is still local code execution.
agentcad executes CadQuery Python scripts and produces STEP files, PNG renders, mesh exports
Use this in a project directory with trusted inputs, review unusual script changes, and install agentcad only from a source you trust.
Installing the skill requires trusting the separate agentcad package and its future updates.
The reviewed artifacts are only the skill manifest and README; the required executable comes from an external PyPI package whose source is not included in this review.
CLI source: private repo, distributed via [PyPI](https://pypi.org/project/agentcad/).
Install agentcad from the official PyPI package, consider pinning a known version, and review the package provenance if using it in sensitive environments.
A browser or viewer may open after builds, and local output files will be generated as part of normal use.
The skill directs the agent to open the generated CAD viewer after successful builds without a separate prompt. This is useful for CAD review but is an automatic local UI action.
agentcad view v1_label/viewer.html # or output.step / output.glb ... Do this every run, unprompted.
If you prefer explicit approval before opening files or browser windows, configure tool approvals or tell the agent not to run the view command automatically.