CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

cuda-translation

Use or modify cuda-rust-wasm and cuda-wasm source translation with a trusted cuda2rust executable, bounded processes, parser tests, and Rust/WGSL validation. Use for CUDA-subset parsing, generated-source correctness, compiler-wrapper integration, or safe memory APIs.

Quality

91%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

90%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

An excellent, dense skill body: every command is executable, boundaries and limits are stated as verifiable numbers, and verification is baked into the workflow with an acceptance gate. The only improvements are an explicit error-recovery loop and moving the buffer-semantics/tiled-attention detail into the reference file.

DimensionReasoningScore

Conciseness

The body contains no explanations of concepts Claude already knows; nearly every sentence carries project-specific contracts Claude cannot infer ("no larger than 1 MiB", "caps nesting at 32", "timeoutMs from 1 to 120,000 ms", "E_INPUT_LIMIT, E_COMPILER_NOT_FOUND..."). Matches anchor 5 — lean, assumes competence, every token earns its place; not anchor 4 because there is no padding to trim.

5 / 5

Actionability

Fully executable guidance throughout: the cargo build command, the complete `node cuda-wasm/cli/index.js transpile` invocation with all flags, a copy-paste-ready `transpileCuda` JS example, and full verification commands (cargo test/check, npm test/build/audit). Covers build, CLI, API, and verification use cases — matches anchor 5.

5 / 5

Workflow Clarity

Clear sequence across sections (Define the unit → Use the checkout compiler → Boundary checklist → Validate generated behavior → Verify → Acceptance and diagnosis) with an explicit checklist and concrete Verify commands. Falls short of anchor 5 because there is no explicit validate-fix-retry feedback loop — error handling states codes and that "a missing compiler is a prerequisite failure" but does not spell out recovery steps.

4 / 5

Progressive Disclosure

Well-sectioned body with a clearly signaled, one-level-deep reference ("Read [translation verification details](references/translation.md) for backend and research boundaries" — the file exists). Minor gap: the HostBuffer/DeviceBuffer and tiled-attention paragraph is deep project detail that arguably belongs in the reference file rather than inline, matching anchor 4 rather than 5.

4 / 5

Total

18

/

20

Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: concrete, third-person, with an explicit 'Use for...' trigger clause covering four specific scenarios and named tools that make it highly distinguishable. The only weaknesses are the slightly generic leading verbs and a few missing natural synonyms like 'transpile' or '.cu'.

DimensionReasoningScore

Specificity

"Use or modify cuda-rust-wasm and cuda-wasm source translation with a trusted cuda2rust executable, bounded processes, parser tests, and Rust/WGSL validation" names the domain and several concrete actions with specific instruments. It sits between anchors 4 and 5: the actions are specific and coverage is broad, but the leading verbs "Use or modify" are generic compared to the fully distinct action enumeration of the anchor-5 example.

4 / 5

Completeness

Explicitly answers both: the what ("source translation with a trusted cuda2rust executable, bounded processes, parser tests, and Rust/WGSL validation") and the when via an explicit trigger clause ("Use for CUDA-subset parsing, generated-source correctness, compiler-wrapper integration, or safe memory APIs") with concrete trigger phrases. Matches anchor 5; the 'when' is explicit and specific, not the weaker implied form of anchor 4.

5 / 5

Trigger Term Quality

Includes good natural keywords — "CUDA-subset parsing", "generated-source correctness", "compiler-wrapper integration", "Rust/WGSL", "safe memory APIs" — that a user of this project would plausibly say. A few natural variations are missing (e.g., "transpile", "convert CUDA to Rust", ".cu files"), keeping it at anchor 4 rather than 5.

4 / 5

Distinctiveness Conflict Risk

A clear niche (CUDA-to-Rust/WGSL source translation for the named cuda-rust-wasm/cuda-wasm/cuda2rust toolchain) with distinct triggers; virtually no other skill would trigger on these terms. Matches anchor 5.

5 / 5

Total

18

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

ruvnet/ruv-FANN

Path

plugins/ruv-fann/skills/cuda-translation/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.