[sambitcreate]

Requested by sbt · project thread

This bumps package.json and package-lock.json to 0.60.0. Merging it to main runs Release desktop, which publishes v0.60.0.

In this release (since 0.59.0)

• Make Bots on one Mac usable from another Mac #560, Harden paired-Mac Bot chats against version skew, and fix the Bot session contract check #565: Bots on one Mac usable from another Mac, plus follow-up fixes.
• Compact on return to a chat whose prompt cache expired #561: compact-on-return for large chats whose prompt cache has expired.
• Steer stops running subagents; draining subagents no longer trap the queue #562: Steer stops running subagents; a stuck cancelled subagent no longer blocks the queue.
• Backport Pi #11142 (Responses prompt cache across tool changes) to pi-ai 1.3.0 #563: Pi #11142 backport (Responses prompt cache across tool changes) and an attended tool-loop recovery stop.
• Small client fixes from T3: Failed rows, typed paths, still MJPEG frames #564: Failed pill, typed paths skip the slash menu, simulator viewer still-screen timeout.
• Open local file links from chat replies in the Aiden browser and Files #566: chat file links open in Aiden Browser/Files, with a right-click choice of browser.
• Fix MCP sign-in and move connectors onto Pi's MCP client #568: MCP audit and move to Pi's MCP client, Connectors settings.
• iOS: fix chat flicker and jitter, and pace streamed text #572: iOS paced streaming and Markdown repair (phones need a new build).
• Instant UI and optimistic updates across desktop Bots, iOS and Android #576: Instant UI and optimistic updates (phones need a new build).
• Improve harness reliability, protected evaluation, and CLI Bot durability #577 (carries Improve harness reliability, protected evaluation, and CLI Bot durability #553): harness reliability foundation, private verification receipts, opt-in CLI durable Bots.

Not included: #559 (Bot hand-offs), which is still a draft.

Docs

• docs/plans/README.md: refreshed for 0.60.0. The MCP, Instant UI and harness rows are marked as shipped.
• docs/aiden-remote-api-v1.md: "After 0.59.0" now reads "From 0.60.0" for the Bot-chat turn refusal.

Flaky E2E fix

• Spec: tests/e2e/browser-throttling.spec.ts:14, the first spec in Electron E2E shard 3/3.
• Symptom: the first attempt failed and the retry passed, so --fail-on-flaky-tests failed the job. The log printed "Downloading Electron binary..." inside the test. The spec then ran into its 60s execFile budget at "starting recording".
• Run: https://github.com/sambitcreate/aiden-agent/actions/runs/39673903473/job/119488872710
• Cause: Electron 43 resolves its binary lazily. On a fresh shard, the first spec paid for the binary download and the first-launch signature check inside its own budget.
• Fix: a Playwright globalSetup (tests/e2e/global-setup.ts) resolves the binary and runs electron --version once before any spec. No timeouts or retries changed.

🤖 Generated with Claude Code

https://claude.ai/code/session_01SE574SfMr5QCRmoYLXy2D7

[pullfrog]

✅ No new issues found.

Reviewed changes Reviewed the complete release-version and documentation diff, including its interaction with the existing desktop release workflow.

• Release identity: package.json and both root lockfile version fields consistently advance to 0.60.0; the release selector produces v0.60.0 with publication enabled when that tag is absent.
• Release documentation: The plan inventory records the merged release contents and outstanding acceptance gates, and the Remote API guide identifies 0.60.0 as the version introducing the existing legacy Bot-turn refusal.
• Validation: All 29 focused release-selection, consumer-contract, publishing, Linux-feed, and CI-policy tests passed. Hosted CI is still running; signed distribution and macOS/Electron acceptance were not run locally on this Linux runner.

  ^{｜ View workflow run ｜ Using gpt-6.1-sol ｜ 𝕏}

[very-hermes-bot]

Hermes Review Bot

Confidence: 5

Engine: agy/gemini-3.8-flash-high
Review mode: full
Head: 3ad7fd3dee2d18d93910d697c950d1d8cf1ba5b1
Generated: 2026-10-21T21:51:30+00:00
Reviews: 1

Summary

Bumps the application version in package.json and package-lock.json from 0.59.0 to 0.60.0 for the desktop release workflow, and updates plan milestones and Remote API documentation to match. It also introduces a Playwright globalSetup hook that invokes the Electron binary with --version once before any specs run, preventing cold-start macOS Gatekeeper signature verification latency on fresh CI shards from exhausting individual test timeouts. Maintainers should double-check that all required CI matrix checks complete cleanly on the head commit prior to merging into main to trigger the automated release build.

Confidence Score: 5/5

5/5: Diff, git history, documentation claims, CI workflows, and the Playwright global setup hook were thoroughly inspected and verified against existing codebase conventions.

📁 Important Files Changed

• package.json & package-lock.json: Declares release version 0.60.0.
• docs/plans/README.md: Records shipped PR milestones (Make Bots on one Mac usable from another Mac #560–Open local file links from chat replies in the Aiden browser and Files #566, Fix MCP sign-in and move connectors onto Pi's MCP client #568, iOS: fix chat flicker and jitter, and pace streamed text #572, Instant UI and optimistic updates across desktop Bots, iOS and Android #576, Improve harness reliability, protected evaluation, and CLI Bot durability #577) and updates the baseline release commit to 32020413.
• docs/aiden-remote-api-v1.md: Updates Bot durable-session turn refusal contract status from "After 0.59.0" to "From 0.60.0".
• playwright.config.ts: Registers globalSetup for Playwright test runs.
• tests/e2e/global-setup.ts: Spawns electron --version with a 120s timeout before specs run to warm binary resolution and Gatekeeper signature checks.

Findings

No findings.

📊 Sequence Diagram

Unchanged or not applicable for this change.

Machine-Readable Findings

[]

---

Last reviewed commit: 3ad7fd3dee2d
Reviews (1) · Comment /hermes review to trigger a new review · /hermes review full for full re-review

[pullfrog]

No new issues found.

Reviewed changes Reviewed the Electron E2E setup changes since the prior review at 4938095, with the complete release diff as context.

• Added suite-level warm-up: Registered a global setup that resolves Electron before the specs and attempts a bounded --version launch, moving lazy binary installation and first-launch work outside individual test budgets.
• Preserved test enforcement: Left spec timeouts, retries, and --fail-on-flaky-tests unchanged; warm-up launch failures remain the specs' responsibility.
• Validated configuration: E2E type-checking passed, Playwright collected 223 tests in 42 files, and all 21 focused sharding and CI-policy tests passed. Direct setup invocation completed on Linux, but macOS first-launch behavior and the original flake remain unverified locally. Exact-head hosted CI was still running.

  ^{｜ View workflow run ｜ Using gpt-6.1-sol ｜ 𝕏}