CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

deepgram-go-voice-agent

Use when writing or reviewing Go code in this repo that runs a Deepgram Voice Agent session over WebSockets, including runtime settings, prompt updates, speak updates, injected messages, and event handling. Route standalone STT/TTS work to deepgram-go-speech-to-text or deepgram-go-text-to-speech.

65

Quality

82%

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

68%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.

A well-structured, token-efficient single-file skill with executable quick-start code and a clearly layered reference list. Its main gap is workflow depth: the session lifecycle's most delicate parts — streaming audio, event handling, the function-call response loop, and reconnect recovery — are named or pointed at rather than sequenced with checkpoints.

Suggestions

Add a short numbered session-lifecycle workflow (open -> settings -> connect -> start -> stream -> handle events -> respond to function calls) with explicit validation checkpoints, e.g. verifying the Welcome event before streaming and how to respond to FunctionCallRequestResponse.

Include minimal executable snippets for the core named operations (e.g. conn.Stream / Write for audio frames, UpdatePrompt, and FunctionCallResponse) instead of listing them only as names.

Trim duplication between the 'When to use this product' section and the frontmatter description, and tighten the closing 'Central product skills' paragraph to a one-line pointer.

DimensionReasoningScore

Conciseness

The body is efficient and assumes competence (no explaining WebSockets or Go basics), but minor padding could be trimmed: the 'When to use this product' section re-derives the frontmatter description, and the closing 'Central product skills' paragraph is semi-promotional. Anchor 4, not 5, because of these trimmable spots and the slightly long quick-start boilerplate.

4 / 5

Actionability

The quick start is real, compilable Go code with correct imports, and concrete in-repo paths are given for payloads ('pkg/api/agent/v1/websocket/interfaces/types.go') and examples. However, streaming, event handling, and UpdatePrompt/UpdateSpeak usage are only named, never shown, so it does not reach anchor 5; the executable quick start keeps it above anchor 3.

4 / 5

Workflow Clarity

A session sequence is present (settings -> handler -> connect -> start -> stream/handle, plus the lifecycle bullets and gotchas like 'Connect() returns bool rather than error'), but the function-call response loop and reconnect/error-recovery checkpoints are only implied, matching anchor 3 ('sequence present but checkpoints missing or implicit') rather than 4.

3 / 5

Progressive Disclosure

As a single-file skill with no bundle files, it is well organized with clearly signaled one-level references (in-repo files, OpenAPI/AsyncAPI, docs) in a layered list. Anchor 4 rather than 5 because the inlined payload/event name list and ~130-line body could partly live in a separate reference file.

4 / 5

Total

15

/

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 with an explicit and specific trigger clause, an enumerated capability list, and proactive disambiguation against sibling skills. Its only weaknesses are minor coverage gaps (audio streaming, function-call handling are not named) and a few missing natural trigger variations.

DimensionReasoningScore

Specificity

The description enumerates several concrete capability areas — 'runtime settings, prompt updates, speak updates, injected messages, and event handling' — plus routing to sibling skills, but omits others the body covers (audio streaming, function-call responses), matching anchor 4 rather than 5.

4 / 5

Completeness

It explicitly answers both questions: a concrete 'what' ('runs a Deepgram Voice Agent session over WebSockets, including runtime settings, prompt updates, speak updates, injected messages, and event handling') and a specific 'when' ('Use when writing or reviewing Go code in this repo...'), matching anchor 5; the 'when' is too explicit and specific for anchor 4.

5 / 5

Trigger Term Quality

Good natural-term coverage ('Go code', 'Deepgram Voice Agent', 'WebSockets', 'prompt updates', 'STT/TTS') including the STT/TTS synonyms, but misses common user variations like 'voice AI', 'real-time voice', or 'transcribe', so it sits at anchor 4 rather than 5.

4 / 5

Distinctiveness Conflict Risk

'Deepgram Voice Agent session over WebSockets' in Go is a clear niche, and the routing sentence ('Route standalone STT/TTS work to deepgram-go-speech-to-text or deepgram-go-text-to-speech') actively steers away from sibling skills, giving minimal conflict risk as in 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

deepgram/deepgram-go-sdk

Path

.agents/skills/deepgram-go-voice-agent/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.