name	gemini-result-handling
description	Internal guidance for presenting Gemini helper output back to the user
user-invocable	false

Gemini Result Handling

When the helper returns Gemini output:

Preserve the helper's verdict, summary, findings, and next steps structure.
For review output, present findings first and keep them ordered by severity.
Use the file paths and line numbers exactly as the helper reports them.
Preserve evidence boundaries. If Gemini marked something as an inference, uncertainty, or follow-up question, keep that distinction.
Preserve output sections when the prompt asked for them, such as observed facts, inferences, open questions, touched files, or next steps.
If there are no findings, say that explicitly and keep the residual-risk note brief.
If Gemini made edits, say so explicitly and list the touched files when the helper provides them.
For gemini:gemini-rescue, do not turn a failed or incomplete Gemini run into a Claude-side implementation attempt. Report the failure and stop.
For gemini:gemini-rescue, if Gemini was never successfully invoked, do not generate a substitute answer at all.
CRITICAL: After presenting review findings, STOP. Do not make any code changes. Do not fix any issues. You MUST explicitly ask the user which issues, if any, they want fixed before touching a single file. Auto-applying fixes from a review is strictly forbidden, even if the fix is obvious.
If the helper reports malformed output or a failed Gemini run, include the most actionable stderr lines and stop there instead of guessing.
If the helper reports that setup or authentication is required, direct the user to /gemini:setup and do not improvise alternate auth flows.

Never fabricate results for incomplete jobs

Never fabricate Gemini output. Do not invent findings, file lists, diffs, or next steps that the helper did not return.
If a run is incomplete, non-terminal, or still running, do not synthesize a final answer. Direct the user to /gemini:status <job-id> for live state, /gemini:result <job-id> once the job completes, or /gemini:cancel <job-id> to stop a runaway job.
For in progress or non-terminal jobs, present only the live snapshot the helper returned and surface the recommended action verbatim.
If the helper reports failed, cancelled, worker_missing, or broker_unhealthy, never fabricate a successful result on top of it — preserve the failure state and recommended action.

Gemini Result Handling

When the helper returns Gemini output:

Preserve the helper's verdict, summary, findings, and next steps structure.

For review output, present findings first and keep them ordered by severity.

Use the file paths and line numbers exactly as the helper reports them.

Preserve evidence boundaries. If Gemini marked something as an inference, uncertainty, or follow-up question, keep that distinction.

Preserve output sections when the prompt asked for them, such as observed facts, inferences, open questions, touched files, or next steps.

If there are no findings, say that explicitly and keep the residual-risk note brief.

If Gemini made edits, say so explicitly and list the touched files when the helper provides them.

For gemini:gemini-rescue, do not turn a failed or incomplete Gemini run into a Claude-side implementation attempt. Report the failure and stop.

For gemini:gemini-rescue, if Gemini was never successfully invoked, do not generate a substitute answer at all.

CRITICAL: After presenting review findings, STOP. Do not make any code changes. Do not fix any issues. You MUST explicitly ask the user which issues, if any, they want fixed before touching a single file. Auto-applying fixes from a review is strictly forbidden, even if the fix is obvious.

If the helper reports malformed output or a failed Gemini run, include the most actionable stderr lines and stop there instead of guessing.

If the helper reports that setup or authentication is required, direct the user to /gemini:setup and do not improvise alternate auth flows.

Never fabricate results for incomplete jobs

Never fabricate Gemini output. Do not invent findings, file lists, diffs, or next steps that the helper did not return.

If a run is incomplete, non-terminal, or still running, do not synthesize a final answer. Direct the user to /gemini:status <job-id> for live state, /gemini:result <job-id> once the job completes, or /gemini:cancel <job-id> to stop a runaway job.

For in progress or non-terminal jobs, present only the live snapshot the helper returned and surface the recommended action verbatim.

If the helper reports failed, cancelled, worker_missing, or broker_unhealthy, never fabricate a successful result on top of it — preserve the failure state and recommended action.

gemini-result-handling