How do I install the Make Op Verify skill?

Run npx skillmds@latest add cvcuda/make-op-verify in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Make Op Verify safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Make Op Verify?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Make Op Verify free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Make Op Verify?

cvcuda (@cvcuda) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Make Op Verify by cvcuda — Agent Skill

Name: Make Op Verify
Author: cvcuda

Make Op Verify

Verify a new CV-CUDA operator against the deterministic final regression checklist (the /make-op done-gate). Use when asked whether a new operator is complete/done, or to gate it before merge - gold reference, bit-exact coverage across the declared support matrix, required layout parity, complement negatives, docs + relnote, benched dtypes, tests that run and pass, and optimization-readiness.

cvcuda 79f3206 1.7 KB Updated Sep 17, 2026

File contents

Make Op — verify

Thin entry point. Run python3 tools/make_op.py <Operator> --phase done --run and interpret per .agents/guidance/MAKE_OP_GUIDELINES.md. This is the strict-regression done-gate: it reuses review-op (all four domains) and optimize-op preflight, then adds the make-op teeth — independent CPU gold reference (COV-GOLD), bit-exact coverage across the declared support matrix (COV-1/COV-CHAN/COV-3/COV-BITEXACT), required equivalent-layout parity (COV-PARITY/COV-5), complement negatives (COV-NEG), every declared dtype benched (COV-2), the always-on NVTX marker present and registered in the runtime marker test (NVTX-1), the operator in the latest relnote (DOC-REL), and tests that compile/run/pass (EXEC-*). Resolve every GAP with its named fix and re-run; resolve each MANUAL at the cited location. Inviolable: never fabricate baselines (GAP [requires CI] → CI regen); never loosen tolerances (bit-exact default; any EXPECT_NEAR in the C++ test is a GAP); layout-support gaps are author work. Findings-first; completion = a re-run shows zero GAP and zero unresolved MANUAL.

Frequently asked questions