whisper

OpenAI's general-purpose speech recognition model. Supports 99 languages, transcription, translation to English, and language identification. Six model sizes from tiny (39M params) to large (1550M params). Use for speech-to-text, podcast transcription, or multilingual audio processing. Best for robust, multilingual ASR.

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

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 code and well-organized sections, but it is padded with redundant metrics/language/resource material and fails to use the bundled references file, leaving language content inline. Batch processing also lacks verification steps.

Suggestions

Replace the inline 'Language support' section with a one-line pointer to references/languages.md to apply progressive disclosure and remove the duplication.

Trim non-instructional padding (GitHub-star Metrics, the Performance real-time-factor table, the Resources list) to tighten conciseness toward level 3.

Add a validation/verification checkpoint to the batch-processing workflow (e.g., confirm each output file was written and is non-empty) so workflow clarity is not capped at 2.

Dimension	Reasoning	Score
Conciseness	Mostly efficient code-driven content, but padded with a GitHub-star 'Metrics' block, a duplicate Performance RTF table, an inline Language Support section that mirrors the unused references file, and a Resources section; not the lean level-3 that assumes Claude's competence.	2 / 3
Actionability	Provides fully executable Python and bash examples (load_model, transcribe with options, CLI flags, batch loop) that are copy-paste ready with specific parameters, matching the level-3 anchor.	3 / 3
Workflow Clarity	Steps are sequenced, but the batch-processing section writes output files with no verification checkpoint, which the rubric caps at 2 for batch operations; it does not reach the explicit validate/retry level 3.	2 / 3
Progressive Disclosure	Section structure is clear, but the provided references/languages.md is never linked and its content is duplicated inline in 'Language support', matching the level-2 anchor where content that should be separate is inline and references are not signaled.	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 concrete, distinct, and gives explicit 'Use for…' trigger guidance covering natural user terms, answering both what and when without verbosity. It is among the strong examples in the rubric.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions ('transcription, translation to English, and language identification') plus use cases (speech-to-text, podcast transcription), matching the level-3 anchor rather than the partial level-2.	3 / 3
Completeness	Answers both what (speech recognition model supporting 99 languages, transcription, translation, language ID) and when via the explicit 'Use for speech-to-text, podcast transcription, or multilingual audio processing' trigger, so it is not capped at 2.	3 / 3
Trigger Term Quality	Natural terms a user would say are well covered ('transcription', 'podcast transcription', 'speech-to-text', 'multilingual audio'); 'ASR' is minor jargon but does not undermine the broad natural coverage, so it is above level 2.	3 / 3
Distinctiveness Conflict Risk	A clearly scoped Whisper/speech-recognition niche with distinct triggers unlikely to conflict with other skills; above the 'could still overlap' level-2 anchor.	3 / 3
	Total	12 / 12 Passed

Validation

93%

Warnings & errors only

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

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/llm-tools/whisper/SKILL.md
Commit: 52845c3

Reviewed: about 13 hours ago

Table of Contents

Discovery Implementation Validation

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.