feat(harness): add a Gemini CLI harness (tap, turn, and init templates)

[michaelxu2288]

What

Adds Gemini CLI as a framework harness, mirroring the Claude Code and Codex ones:

• convert_gemini_cli_to_agentex_events (src/agentex/lib/adk/_modules/_gemini_cli_sync.py): maps the CLI's stream-json events (init, message deltas, tool_use, tool_result, error, result; schema from packages/core/src/output/types.ts in google-gemini/gemini-cli) onto the canonical StreamTaskMessage* stream. Assistant deltas open one text slot that closes on the next tool event, the result, or end of stream, so every Start has a Done; tool requests and results pair by tool_id; a finally closes the source iterator on cancellation like the other taps.
• GeminiCliTurn (_gemini_cli_turn.py): HarnessTurn wrapper exposing session_id and model from init and normalising result.stats into TurnUsage (tokens, cached, duration, tool calls; cost is not reported by the CLI).
• Both exported from agentex.lib.adk.
• agentex init templates sync-gemini-cli, default-gemini-cli, temporal-gemini-cli, registered in TemplateType, the file map and the menus. Prompt passed via -p with stdin closed (in headless mode the CLI reads stdin to EOF and appends it), optional GEMINI_MODEL, GEMINI_API_KEY credential, npm install -g @google/gemini-cli in the Dockerfile. Turns are independent prompts because the CLI's --resume takes latest/index rather than a session id; the Temporal template keeps the reported session_id for observability only and says so.

Companion docs page: scaleapi/scale-agentex#429.

Test

• tests/lib/adk/test_gemini_cli_sync.py (12), tests/lib/adk/test_gemini_cli_turn.py (12), tests/lib/core/harness/test_harness_gemini_cli_sync.py (5): pass.
• tests/lib/cli/test_init_templates.py: 29 passed, including the three new templates; all rendered projects byte-compile.
• ruff check and format clean on the new files.
• Offline fixtures follow the CLI's published event schema; I have not run a live smoke test against the gemini binary (needs a Gemini API key). Happy to add a recorded live fixture if a maintainer can share one, or run it myself once I have a key.

Confidence Score: 4/5

The PR is not ready to merge because a failed Gemini turn can still finish as a successful turn.

Findings

1.  Terminal failures appear successful ▶

Fix with agent prompt

### Issue 1
src/agentex/lib/adk/_modules/_gemini_cli_sync.py:281-291
A terminal Gemini `result` can have `status: "error"`, but this branch closes the text stream and invokes the normal result callback without checking the status or error payload. `UnifiedEmitter` therefore treats authentication, turn-limit, and other terminal failures as successful completion, potentially returning missing or partial output instead of surfacing the error. Standalone error events are also logged without entering the canonical stream.

---

For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.

Summary

Adds Gemini CLI support to the Agentex harness and agentex init, so developers can build Gemini CLI agents with sync, async, or Temporal delivery.

• Converts Gemini CLI events and usage stats into Agentex harness messages.
• Adds starter projects for all three delivery styles, including CLI setup and credentials.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  A[Gemini CLI output] --> B[GeminiCliTurn]
  B --> C[Task-message events]
  C --> D[Sync response]
  C --> E[Async task stream]
  C --> F[Temporal activity]

Loading

_{Reviews (3) · Last reviewed commit: "test(harness): narrow content types in t..."}

[greptile-apps]

Terminal failures appear successful

A terminal Gemini result can have status: "error", but this branch closes the text stream and invokes the normal result callback without checking the status or error payload. UnifiedEmitter therefore treats authentication, turn-limit, and other terminal failures as successful completion, potentially returning missing or partial output instead of surfacing the error. Standalone error events are also logged without entering the canonical stream.

Knowledge Base Used:

• Agent framework integrations
• Harness delivery and metrics

Prompt To Fix With AI

This is a comment left during a code review.
Path: src/agentex/lib/adk/_modules/_gemini_cli_sync.py
Line: 262-267

Comment:
**Terminal failures appear successful**

A terminal Gemini `result` can have `status: "error"`, but this branch closes the text stream and invokes the normal result callback without checking the status or error payload. `UnifiedEmitter` therefore treats authentication, turn-limit, and other terminal failures as successful completion, potentially returning missing or partial output instead of surfacing the error. Standalone error events are also logged without entering the canonical stream.

**Knowledge Base Used:**
- [Agent framework integrations](https://app.greptile.com/scale-ai/-/custom-context/knowledge-base/scaleapi/scale-agentex-python/-/docs/agent-framework-integrations.md)
- [Harness delivery and metrics](https://app.greptile.com/scale-ai/-/custom-context/knowledge-base/scaleapi/scale-agentex-python/-/docs/harness-and-metrics.md)

---

For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.

[michaelxu2288]

Fixed in d9b1ceb. A result with status: "error" now emits a TextContent error ("Error: Gemini CLI turn failed: ") before on_result fires, the same shape the Codex tap uses for turn.failed. Standalone error events with severity "error" are delivered the same way; warnings are still only logged. Covered by tests at the tap level and through UnifiedEmitter.yield_turn.