feat(python): voice-agents v1 — conversation-memory

[github-actions]

Adds the conversation-memory recipe for voice-agents v1 (python).

Demonstrates two patterns for maintaining conversation context:

• Initial context: pre-loads conversation history via AgentV1SettingsAgentContext
• Dynamic prompt updates: tracks ConversationText events and uses send_update_prompt() to inject a sliding window of recent turns into the system prompt

Closes part of #533

[claude]

@deepgram/devrel — VP escalation: PR #538 has had no CI checks and no reviews for >48 hours. Workflow dispatch failed (GITHUB_TOKEN 403). Manual review or test trigger needed.

State: type:samples, no checks, no reviews, idle >48h

[claude]

@deepgram/devrel — VP escalation: PR #538 has been stuck for >4 days with no CI checks running and no review. Lead — Test workflow dispatch failed (GITHUB_TOKEN limitation). Manual re-trigger or review needed.

State: labels=[type:samples, language:python], no checks, no reviews.

[claude]

@deepgram/devrel — VP escalation: PR #538 has been stuck for >4 hours with no CI checks running and re-trigger attempts failed (GITHUB_TOKEN permission). Manual review needed.

State: No review, no passing checks. Workflow dispatch blocked by token permissions.

[claude]

@deepgram/devrel — VP escalation: PR #538 has no CI checks and has been open >24h. The VP agent cannot dispatch workflows (GITHUB_TOKEN 403). Manual re-trigger of CI needed.

[claude]

@deepgram/devrel — VP escalation: PR #538 has had no CI checks run. The Lead — Test workflow was not triggered (likely due to GITHUB_TOKEN event propagation limitation). Manual re-trigger of Lead — Test workflow needed.

State: type:samples label, zero checks, last updated 2026-05-05