Skip to content
DevAgentRadarDevAgentRadar
Sign in
Proposals/Upgrade Gemini CLI to v0.56.0-preview.1 to end f
proposalteamP5Worth a lookGCGemini CLI

Upgrade Gemini CLI to v0.56.0-preview.1 to end false capacity exhaustion

Move to Gemini CLI v0.56.0-preview.1 so false capacity-exhaustion failures stop, real capacity exhaustion is terminal, MCP OAuth refreshes with the stored client ID, and the new local eval report path is available.

Why this loop

This preview fixes false model capacity exhaustion and core quota lookup model mapping, and reclassifies capacity exhaustion as a terminal error so agents fail fast instead of thrashing on bad quota signals. It also refreshes MCP OAuth tokens with the stored client ID, which reduces broken MCP tool sessions after expiry. feat(evals) adds a local report command plus developer documentation, giving a concrete loop for prompt/tool regressions. Upgrade if your runners show spurious capacity errors, flaky MCP auth, or missing local eval reporting.

Proposed actions

  1. Upgrade Gemini CLI to exactly v0.56.0-preview.1 from https://github.com/google-gemini/gemini-cli/releases/tag/v0.56.0-preview.1 and pin that version in the team agent runner image or devcontainer.
  2. Reconnect every MCP server that uses OAuth after the upgrade so token refresh uses the stored client ID fix (PR #28481).
  3. Replay a transcript that previously hit false capacity exhaustion; confirm the quota lookup model mapping fix stops the false error and true capacity exhaustion exits as terminal.
  4. Using the evals developer documentation added in this release, run the new local report command on your golden prompt set and save the report next to the prompt pack.
  5. Update runner wrappers: do not auto-retry capacity-exhaustion errors now that they are classified as terminal.

Agent prompt

Paste into your agent or query via MCP (`get_agent_prompt`) — free, no extra AI cost

Google Antigravity / agent task

# DevAgentRadar → Google Antigravity

Goal: turn this release signal into a safe pilot plan for my stack.

## Context
Assistant: Gemini CLI
Proposal: Upgrade Gemini CLI to v0.56.0-preview.1 to end false capacity exhaustion
Summary: Move to Gemini CLI v0.56.0-preview.1 so false capacity-exhaustion failures stop, real capacity exhaustion is terminal, MCP OAuth refreshes with the stored client ID, and the new local eval report path is available.
Primary source: https://github.com/google-gemini/gemini-cli/releases/tag/v0.56.0-preview.1

## Why it matters
This preview fixes false model capacity exhaustion and core quota lookup model mapping, and reclassifies capacity exhaustion as a terminal error so agents fail fast instead of thrashing on bad quota signals. It also refreshes MCP OAuth tokens with the stored client ID, which reduces broken MCP tool sessions after expiry. feat(evals) adds a local report command plus developer documentation, giving a concrete loop for prompt/tool regressions. Upgrade if your runners show spurious capacity errors, flaky MCP auth, or missing local eval reporting.

## Suggested actions
1. Upgrade Gemini CLI to exactly v0.56.0-preview.1 from https://github.com/google-gemini/gemini-cli/releases/tag/v0.56.0-preview.1 and pin that version in the team agent runner image or devcontainer.
2. Reconnect every MCP server that uses OAuth after the upgrade so token refresh uses the stored client ID fix (PR #28481).
3. Replay a transcript that previously hit false capacity exhaustion; confirm the quota lookup model mapping fix stops the false error and true capacity exhaustion exits as terminal.
4. Using the evals developer documentation added in this release, run the new local report command on your golden prompt set and save the report next to the prompt pack.
5. Update runner wrappers: do not auto-retry capacity-exhaustion errors now that they are classified as terminal.

## Config surfaces this release may change
- MCP servers — check your repo before applying


Please:
1. Map the change to concrete pilot steps
2. Flag security / permission implications
3. Keep the pilot reversible
agentmcpmodelcliprereleaseRelease source ↗

Your loop

This browser · no sign-in · not shared as “you”

Your decision stays on this device. A public tally appears after a few votes.

Originating release signal

GCGemini CLIv0.56.0-preview.1Aug 11, 2026

v0.56.0-preview.1 Release v0.56.0-preview.1

Changelog for v0.55.0-preview.1 · chore(release): bump version to 0.56.0-nightly.20260806.g761f604c1 · Changelog for v0.54.0 · Reclassifying Capacity Exhaustion as Terminal Error · feat(caretaker): update Firestore schema with error, and pr_number fields · feat(caretaker-triage): prompt hill-climbing & orchestrator updates · feat(caretaker): add triage Cloud Run job workflow · feat(caretaker-evals): add triage evaluation framework and judge runner · feat(caretaker-evals): add local golden issue collection and firestore sync tools · +8 more changes

Verified excerpt — the source's own words
## What's Changed
* Changelog for v0.55.0-preview.1 by @gemini-cli-robot in https://github.com/google-gemini/gemini-cli/pull/28706
* chore(release): bump version to 0.56.0-nightly.20260806.g761f604c1 by @gemini-cli-robot in https://github.com/google-gemini/gemini-cli/pull/28707
* Changelog for v0.54.0 by @gemini-cli-robot in https://github.com/google-gemini/gemini-cli/pull/28708
* Reclassifying Capacity Exhaustion as Terminal Error by @luisfelipe-alt in https://github.com/google-gemini/gemini-cli/pull/28716
* feat(caretaker): update Firestore schema with error, and pr_number fields by @chadd28 in https://github.com/google-gemini/gemini-cli/pull/28467
* feat(caretaker-triage): prompt hill-climbing & orchestrator updates by @chadd28 in https://github.com/google-gemini/gemini-cli/pull/28524
* feat(caretaker): add triage Cloud Run job workflow by @chadd28 in https://github.com/google-gemini/gemini-cli/pull/28468
* feat(caretaker-evals): add triage evaluation framework and judge runner by @chadd28 in https://github.com/google-gemini/gemini-cli/pull/28530

Excerpt ends here — this release continues at the source ↗.

Primary source ↗