deepgram-core-workflow-b

Implement real-time streaming transcription with Deepgram WebSocket. Use when building live transcription, voice interfaces, real-time captioning, or voice AI applications. Trigger: "deepgram streaming", "real-time transcription", "live transcription", "websocket transcription", "voice streaming", "deepgram live".

Quality

83%

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

77%Weight 40%Scale 1-3

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

A highly actionable, well-sequenced skill body with executable code and solid error-recovery guidance. Its main gaps are token efficiency (repeated option boilerplate) and progressive disclosure: the existing reference file is unreferenced and overlaps the inline content.

Suggestions

Link `references/streaming-implementation.md` from the body (e.g. a 'See [streaming-implementation.md](references/streaming-implementation.md) for a consolidated manager class and SSE example' line) and move the duplicated setup/diarization/SSE detail there to reduce inline repetition.

Extract the repeated `listen.live({ ... })` option block into a single base config snippet referenced by Steps 1, 3, and 5 to tighten token usage.

Reconcile the model mismatch (body uses `nova-3`, the reference uses `nova-2`) so the two files stay consistent.

Dimension	Reasoning	Score
Conciseness	The prose is lean and avoids explaining concepts Claude already knows, but the same `deepgram.listen.live({ ... })` option block is repeated nearly verbatim across Steps 1, 3, and 5, and inline option comments restate what the field names already convey — it could be tightened by referencing a shared base config.	2 / 3
Actionability	Every step ships complete, copy-paste-ready TypeScript with real imports, exact Sox flags, concrete install commands, and an issue/cause/solution error table — fully executable rather than pseudocode.	3 / 3
Workflow Clarity	Steps are numbered and sequenced (basic transcription -> mic capture -> diarization -> auto-reconnect -> SSE -> keepalive) with explicit error-recovery feedback loops (Close/Error handlers, backoff reconnection, KeepAlive on idle timeout) and a graceful SIGINT shutdown path.	3 / 3
Progressive Disclosure	The body is well sectioned, but the provided bundle file `references/streaming-implementation.md` is never linked or signaled from the body, and its content (live setup, mic capture, diarization, SSE endpoint) substantially duplicates the inline steps — material that should live in the reference is inline and the reference is not navigable.	2 / 3
	Total	10 / 12 Passed

Description

90%Weight 40%Scale 1-3

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, well-structured description with explicit 'Use when' guidance and a dedicated natural-language trigger list. Its only weakness is specificity: it leads with a single action verb rather than listing the multiple concrete capabilities the body actually covers.

Suggestions

Expand the opening to list concrete actions (e.g. 'Capture microphone audio, handle interim/final results, diarize speakers, auto-reconnect, and expose an SSE endpoint') so specificity reaches the score-3 anchor.

Drop redundant trigger phrasing that overlaps the 'Use when' clause to keep the description tight.

Dimension	Reasoning	Score
Specificity	Names the domain and one concrete action ('Implement real-time streaming transcription with Deepgram WebSocket'), but does not enumerate multiple distinct capabilities the way the score-3 anchor does; it is a single verb rather than a comprehensive action list.	2 / 3
Completeness	It explicitly answers both 'what' ('Implement real-time streaming transcription with Deepgram WebSocket') and 'when' ('Use when building live transcription, voice interfaces, real-time captioning, or voice AI applications' plus the Trigger line), matching the score-3 anchor that requires explicit triggers.	3 / 3
Trigger Term Quality	The explicit 'Trigger:' line supplies six natural phrases a user would actually say ('deepgram streaming', 'real-time transcription', 'live transcription', 'websocket transcription', 'voice streaming', 'deepgram live'), giving good coverage of common variations.	3 / 3
Distinctiveness Conflict Risk	The Deepgram WebSocket streaming niche is specific with distinct Deepgram-scoped triggers, making it unlikely to fire for unrelated transcription or file-processing skills.	3 / 3
	Total	11 / 12 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: jeremylongshore/claude-code-plugins-plus-skills
Path: plugins/saas-packs/deepgram-pack/skills/deepgram-core-workflow-b/SKILL.md
Commit: 4c47e33

Reviewed: 1 day 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.