agent-all-gemini

/agent-all (Gemini port)

Runs the cost-unrestricted multi-agent pipeline using Gemini CLI primitives. Phase 3 dispatch is subprocess-based: each wave task spawns its own gemini chat subprocess via run_shell_command, all in parallel. The coordinator awaits via process exit + output file parsing.

Usage

/agent-all-gemini "add user signup form"
/agent-all-gemini docs/tasks/12-fix-login.md
/agent-all-gemini "fix flaky test" --loop --max-iter=5
/agent-all-gemini docs/tasks/x.md --no-pr --wave-size=large

Flags

Same as Claude Code: --loop, --max-iter=<N>, --max-cost=<USD>, --wave-size=small|medium|large, --no-pr, --no-brainstorm, --resume, --force, --yes, --break-condition=<spec>, --reconfigure.

When --loop is set, Phase 0 prompts the user interactively (via Gemini's ask_user) for the break-condition preset (test-auto / visual-qa / Custom shell / Composite) and offers to save the choice to .agent-all.json. Use --break-condition=<spec> to skip the prompt for one invocation, or --reconfigure to re-prompt even when a non-default value already lives in config.

Additional Gemini-specific:

• --subprocess-timeout=<seconds> — per-task subprocess timeout (default 1800).
• --max-subprocesses=<N> — additional clamp on wave maxParallel (default 8).

Pipeline

Phase	File	Purpose
0	phases/0-preflight.md	git + .gemini/settings.json + subprocess sanity + config
1	phases/1-intent.md	brainstorm (chat) OR load task file
2	phases/2-plan.md	draft plan via write_file
3	phases/3-dispatch.md	fork N parallel gemini chat subprocesses per wave
4	phases/4-gate.md	spawn reviewer subprocesses (same pattern)
5	phases/5-pr.md	run_shell_command: git branch push + gh pr create
6	phases/6-loop.md	breakCondition shell + state.iter re-entry

Rules

1. You orchestrate; phases are source of truth. Read each phase file before running it.
2. State lives in .agent-all-state.json. Atomic write via write_file (write to .tmp then run_shell_command rename).
3. Delegate, don't reimplement. Phase 3 spawns subprocesses; Phase 4 same; Phase 5 uses run_shell_command.
4. Loop is opt-in. Without --loop, Phase 6 is a no-op.
5. Hard caps: --max-iter clamped to 50; --max-cost enforced by reading per-subprocess token logs (best-effort).

Gemini primitive map

Action	Gemini primitive
Read file	read_file
Write file	write_file
Shell	run_shell_command
Dispatch subagent	spawn gemini chat -p "<prompt>" --output-json subprocess
Await dispatched agent	subprocess exit + read its stdout JSON
Persist plan/state	write_file + atomic rename via run_shell_command
Prompt user	ask_user
Invoke another skill in same process	activate_skill

On error

Same as Claude port. Additional Gemini-specific notes:

• If gemini not in PATH: abort gemini binary required for subprocess dispatch.
• If subprocess times out (per --subprocess-timeout): kill it, mark task failed, continue wave.
• If --max-subprocesses reached: queue extra tasks, dispatch as slots free.
• If write_file fails (disk full, permissions): abort state persistence failed.

When done

Print summary: phases completed, iters, cost, PR URL, max parallel subprocesses used. Exit 0/1/2/3 per spec.

References

• references/porting-notes.md — subprocess strategy rationale, alternatives considered
• plugins/harness-floor/skills/agent-all/SKILL.md — source-of-truth pipeline