CDC & Structural Analysis Tool Profiles

What it does

Provides curated tool profiles, conventions, and guidance for Clock Domain Crossing (CDC) and structural analysis workflows. It explains how to run open-source svlens modes (cdc, conn, metrics, all), how to interpret key JSON outputs (violations, cautions, scores), and how to integrate commercial tool runners (SpyGlass, VC_CDC, Questa) into a CI gate. The skill emphasises a quantitative+qualitative gate philosophy: automated svlens reports are paired with LLM judgment and documented waivers where needed.

When to use it

Use this skill when adding CDC/structural checks to CI pipelines, writing runners that normalise disparate tool outputs into a common JSON schema, or when building LLM-assisted verification gates that need both numeric metrics and human/LLM review. It's targeted at verification engineers integrating svlens or commercial CDC runners into lint and gating scripts.

What's included

• Scripts: recommends a wrapper lint/scripts/run_cdc.sh (not included in the SKILL.md payload) and describes typical command lines for svlens and commercial tools.
• References: links to the svlens repo and gives expected JSON fields like summary.violations, analysis.overall_score, and svlens_summary.json.
• Instructions: explains modes (svlens cdc, conn, metrics, all), suggested outputs, interpretation heuristics (e.g. logic_depth_est thresholds), and gate rules (fail on unwaived VIOLATION).

Compatible agents

Best used by engineer-facing automation agents and CI integrators that can run shell commands and parse JSON (Copilot/Copilot-like automation, CLI-capable agents).