CtrlK
BlogDocsLog inGet started
Tessl Logo

evaluating-code-models

Evaluates code generation models across HumanEval, MBPP, MultiPL-E, and 15+ benchmarks with pass@k metrics. Use when benchmarking code models, comparing coding abilities, testing multi-language support, or measuring code generation quality. Industry standard from BigCode Project used by HuggingFace leaderboards.

68

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

65%

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

The body is highly actionable with executable commands and scripts and well-sequenced workflow checklists, but it is verbose with redundant command variants and keeps reference-style material inline instead of linking to the existing reference files. Batch workflows also lack inline validation feedback loops.

Suggestions

Link to the existing references from the body (e.g., 'For the full benchmark list see references/benchmarks.md; for troubleshooting see references/issues.md') and move the Supported Benchmarks and Command Reference tables there to slim the body.

Add an explicit validation/verification checkpoint inside each evaluation workflow (e.g., after Step 3, confirm the results JSON exists and pass@k values are non-null before analyzing) with a fix→retry loop rather than relying only on the Common Issues section.

Deduplicate the repeated accelerate-launch command blocks by defining the common flags once and showing only the per-workflow deltas.

DimensionReasoningScore

Conciseness

The ~400-line body is mostly executable guidance (good) but padded with repeated full accelerate-launch command blocks across workflows, a large command-reference table, and a hardware table that earn their place less than the core steps; it is efficient in places but could be tightened and moved out, matching the 'mostly efficient but could be tightened' anchor rather than the lean score-3 example.

2 / 3

Actionability

Provides fully executable copy-paste commands (accelerate launch invocations with all flags), a runnable bash eval loop, and a Python comparison script, plus a command reference with defaults — matching the score-3 'fully executable, copy-paste ready' anchor.

3 / 3

Workflow Clarity

Each workflow has an explicit checklist and numbered steps (sequence present), but batch code-execution operations lack inline validation checkpoints or fix→retry feedback loops — errors are deferred to a separate 'Common Issues' troubleshooting section rather than validated within the flow, which the rubric caps at 2 for batch operations.

2 / 3

Progressive Disclosure

Three reference files exist (benchmarks.md, custom-tasks.md, issues.md) but the body never links to or signals them, while large reference-style content (the Supported Benchmarks and Command Reference tables) stays inline — matching the score-2 anchor of structure with references present but not clearly signaled and content that should be separate kept inline.

2 / 3

Total

9

/

12

Passed

Description

100%

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

The description is specific, complete, and uses third-person voice with an explicit 'Use when' trigger clause naming multiple natural user phrases. It clearly distinguishes the skill and avoids vague fluff.

DimensionReasoningScore

Specificity

Lists multiple specific concrete actions — 'Evaluates code generation models across HumanEval, MBPP, MultiPL-E, and 15+ benchmarks with pass@k metrics' plus 'benchmarking... comparing coding abilities, testing multi-language support... measuring code generation quality' — matching the score-3 anchor of multiple specific actions rather than a single domain mention.

3 / 3

Completeness

Explicitly answers both 'what' (evaluates code generation models across benchmarks with pass@k) and 'when' via the 'Use when...' trigger clause, matching the score-3 good example structure.

3 / 3

Trigger Term Quality

Covers natural phrases a user would actually say ('benchmarking code models', 'comparing coding abilities', 'testing multi-language support', 'measuring code generation quality'), which is good coverage of common variations rather than only technical jargon.

3 / 3

Distinctiveness Conflict Risk

It carves a clear niche (BigCode code-model benchmarking with pass@k metrics) and names the BigCode Project/HuggingFace leaderboards, with triggers unlikely to fire for unrelated skills.

3 / 3

Total

12

/

12

Passed

Validation

93%

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

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository
foryourhealth111-pixel/Vibe-Skills
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.