NVIDIA / NVIDIA/NemoClaw

[R1a] Define and implement correlated VoiceClaw submission receipts and speech/detail results

Open
#11,753 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
TypeScript
Stars
22.5k
Forks
3.1k
Avg merge
1d 1h
Merged PRs (30d)
715

Description

## Outcome

One finalized client request has a stable identity, explicit acceptance, and matching agent-authored speech/detail with a correlated terminal outcome.

## Scope and ownership

R1a operational follow-up on the v1 deployment established by R0. This issue is not an R0 completion dependency. Its bounded experimental profile must list remaining durable/replay requirements before full contract §9 qualification.

Proposed owner: NemoClaw agent-integration and VoiceClaw request/output owners.

This ticket records requested work. New product boundaries and supported combinations require the owning maintainer decision; no existing proposal is declared accepted here.

## Work

- Extend the public Agent Session API subset established by #11749 with finalized-text submission, status, correlated receipts and semantic result events. Native/ACP translation stays private inside NemoClaw.
- VoiceClaw supplies commandId for idempotent submission. NemoClaw owns attachment/session/task IDs and acceptance/completion receipts. Keep native dispatch evidence distinct from the NemoClaw acceptance receipt and keep native identifiers private.
- Agree a result payload carrying concise agent-authored speech plus associated detail/source references; preserve meaning across the VoiceClaw client interface.
- Record acknowledgement, native acceptance, substantive output and audio-delivery timing at the owning boundaries.
- Version the R1a contract additions with VoiceClaw and run deterministic examples against both implementations.

## Acceptance

- [ ] One finalized request reaches the intended real agent and yields a NemoClaw-issued acceptance receipt, public session/task identity and correlated terminal outcome.
- [ ] The client receives matching concise speech text and detail/source content through documented artifacts.
- [ ] Stable normalized errors carry correlation; oversized input is rejected before dispatch and malformed native output is quarantined.
- [ ] Speech/detail output preserves agent-authored meaning. VoiceClaw owns timing, safe rendering, redaction and delivery; audio and media signaling do not cross into NemoClaw.
- [ ] The tests do not claim durable work or replay across process restart.

## Dependencies

Related R0 epic: [#11746](https://github.com/NVIDIA/NemoClaw/issues/11746). This follow-up is outside R0 completion.

VoiceClaw counterpart: [VoiceClaw #7](https://gitlab-master.nvidia.com/jarvis/voice-claw/-/issues/7).

Blocked by:

- [#11747: [R0] Agree the VoiceClaw installation and attachment contract with conformance fixtures](https://github.com/NVIDIA/NemoClaw/issues/11747)
- [#11749: [R0] Connect VoiceClaw to one v1 agent with a scoped credential](https://github.com/NVIDIA/NemoClaw/issues/11749)

## Category

Feature

Contributor guide

Open the contributing guide

Research direction

No implementation files or tests are named. Start by reading the established API subset in #11749, the conformance work in #11747, and the VoiceClaw counterpart issue #7; then resolve the maintainer decisions and contract boundaries before implementation. Done means the listed acceptance scenarios are covered by deterministic examples without claiming restart durability or replay.

Written by the indexing model from the issue text.

Assessment

Tech stack
typescript
Domain
api, backend-api-design
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.