Skip to content

feat(runtimes): implement multimodal I/O for each CLI runtime #386

@alexey-pelykh

Description

@alexey-pelykh

Tracking issue

Per-runtime multimodal I/O implementation, decomposed into individual issues. This issue tracks the overall effort — implementation details are in the per-runtime issues below.

Depends on #385 (AgentRuntime multimodal contract, done ✅).

Issue Runtime Media support Phase
#397 Gemini images, audio, video, PDF via `@path` Phase 2
#396 Claude images via `--input-format stream-json` stdin Phase 4
#398 Codex blocked upstream (codex#5773) out of scope
#399 OpenCode blocked upstream (hardcoded text/plain MIME) out of scope

All runtimes declare `emitsOutbound: false` (no native media emission).

Related

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions