evanluo42/bench-creator

Maintain a private global AI-work benchmark across repositories by checkpointing code before AI mutation, harvesting reusable difficulties afterward, validating privacy and capture policy, and running candidates in restored evaluator-isolated workspaces. Use before the first file mutation in a substantive repository task when an initialized global Bench enables auto or suggest capture; use after surprising failures, hidden constraints, fragile integrations, or specialized verification; and use whenever the user asks to initialize, register, capture, curate, validate, export, promote, or run Bench cases. Never use a solved live workspace as candidate input.

Claude Skills are markdown + scripts loaded into Claude Code; scanned with the same pipeline.

Source:

github.com/evanluo42/bench-creator

Provenance:

Found by crawling (provenance, not a security rating)

We found this by crawling public sources and nobody has attested to it. That is how most of the registry gets here, and it is not a mark against the artifact — it only means no authenticated act of publication is on record. An artifact can be the vendor's own and still appear here.

1 critical finding reported by cisco-skill-scanner. The composite score alone does not reflect this — read the per-scanner reports below.