RTL CDC Linter
Looks for unsafe clock domain crossings before formal CDC tools run.
Each skill does one clear engineering job and returns one structured output. The current repo has RTL review skills and DV planning skills, not generic "do everything" agents.
Looks for unsafe clock domain crossings before formal CDC tools run.
Estimates combinational depth on register-to-register paths before synthesis.
Turns a short design brief and RTL context into stable DV targets.
Maps the visible parts of the block that matter for verification.
Picks the smallest UVM environment and test list that can verify the stated objectives.
Suggests conservative assertion candidates that match visible block behavior.
Creates deterministic coverpoints, crosses, and exclusions tied to explicit objectives.
Merges the intermediate planning artifacts into one final structured DV plan.
These skills are intentionally narrow. Most of them create reports or plans, not runnable UVM code, signoff results, or full-chip verification architecture.