Skip to content

Releases: QuanFeel/multica-agent-sync

v1.2.1

Choose a tag to compare

@QuanFeel QuanFeel released this 01 Sep 01:52
d45713b

Fixed

  • Select the current authentication config family before considering legacy compatibility configuration, so stale credentials cannot override the active login.
  • Make /multica doctor verify the selected token through /api/me and distinguish ready, invalid, unavailable, and missing states.
  • Redact configuration-read and unexpected diagnostic failures so user-visible Hook output never exposes local paths, credentials, or raw stderr.

Verification

  • CI passed on the release pull request.
  • ./scripts/test.sh: 46/46 tests passed.
  • ./scripts/smoke-install.sh .: isolated Codex plugin installation passed for version 1.2.1.
  • Independent QA review: approved.

v1.2.0

Choose a tag to compare

@QuanFeel QuanFeel released this 21 Aug 11:32

1.2.0 - 2026-08-21

Added

  • Bundle one /multica-sync action Skill for Codex integration while
    keeping /multica commands as the recommended user entry point.

Changed

  • Use main as the repository's only long-lived branch and remove the
    develop release channel from contributor and installation guidance.

Fixed

  • Detect when a Codex task is already tracking a different issue and show an
    explicit, non-automatic rebind sequence instead of injecting the newly
    requested issue context.
  • Return the stop confirmation through the same visible UserPromptSubmit result
    path as the other informational chat commands.
  • Scope status and rebind checks to the exact current Codex task, and refuse
    task-scoped commands when the task id is unavailable.

Multica Agent Sync 1.1.4

Choose a tag to compare

@QuanFeel QuanFeel released this 20 Jul 08:29
05d0207

Changed

  • Removed the bundled Help, Doctor, Status, and Stop Skills. They no longer appear in the Skill picker or start a model-driven Skill turn.
  • /multica help, /multica doctor, /multica status, and /multica stop remain direct UserPromptSubmit Hook controls and block before the prompt reaches the model.
  • /multica 4158 still starts tracking and intentionally continues with exact Multica issue context so Codex can perform the bound task.

Upgrade

  1. Open Settings → Plugins → Marketplaces and click Upgrade for Multica Agent Sync.
  2. Confirm plugin version 1.1.4.
  3. Fully quit and reopen Codex Desktop.
  4. Review and Trust the modified Hook in Settings → Hooks if requested.
  5. Start a new task. Old Skill entries disappear after the restart/new-task boundary.

The Upgrade button remains a persistent manual refresh action; verify the installed version instead.

Multica Agent Sync 1.1.3

Choose a tag to compare

@QuanFeel QuanFeel released this 20 Jul 08:14
f8702ff

Fixed

  • A stale UserPromptSubmit Hook now exits successfully when an upgrade has removed its old versioned plugin cache.
  • Ordinary messages are no longer blocked in every Codex task during the restart window after future upgrades.

Why Upgrade remains visible

The marketplace Upgrade button is a manual refresh action, not an update-available indicator. It remains visible even when the installed marketplace is current. Verify the plugin version instead.

Recovery from 1.1.2

  1. Fully quit and reopen Codex Desktop first to clear the stale 1.1.1 Hook currently held in memory.
  2. Open Settings → Plugins → Marketplaces and click Upgrade for Multica Agent Sync.
  3. Confirm version 1.1.3.
  4. Fully quit and reopen Codex Desktop again before sending messages.
  5. Review and Trust the modified Hook in Settings → Hooks, then start a new task.

This one-time restart sequence is required when upgrading from an older unguarded Hook. Version 1.1.3 and later fail open if a future upgrade replaces their cache before restart.

Multica Agent Sync 1.1.2

Choose a tag to compare

@QuanFeel QuanFeel released this 20 Jul 08:02
84b4f63

Maintenance release

This release exists to validate Codex Desktop manual marketplace Upgrade. Runtime behavior, Hook commands, and the Help/Doctor/Status/Stop Skills are unchanged from 1.1.1.

To test:

  1. Open Settings → Plugins → Marketplaces.
  2. Find Multica Agent Sync and click Upgrade.
  3. Confirm the installed plugin version changes from 1.1.1 to 1.1.2.
  4. Fully restart Codex Desktop and start a new task.

No Hook re-Trust should be required because the Hook definition is unchanged; re-Trust only if Codex marks it modified.

Multica Agent Sync 1.1.1

Choose a tag to compare

@QuanFeel QuanFeel released this 20 Jul 07:55
19453e7

Fixed

  • Restore Help, Doctor, Status, and Stop in the Codex Skill picker.
  • Keep all four Skills explicit-only so they do not trigger automatically.
  • Keep direct /multica help|doctor|status|stop commands available.
  • Do not package Cleanup or Multica Plugin Development.

After upgrading, fully restart Codex Desktop and start a new task so the Skill picker reloads.

Multica Agent Sync 1.1.0

Choose a tag to compare

@QuanFeel QuanFeel released this 20 Jul 07:41
463c657

Highlights

  • Use /multica help and /multica doctor directly in Codex Desktop.
  • Remove all runtime Skills from the public plugin, freeing four Skill list entries.
  • Keep cleanup and purge out of chat and the Skill picker.
  • Add Codex Desktop marketplace Upgrade instructions.

Security

Doctor output is redacted and does not expose Multica credentials or private local paths. Cleanup remains an internal, ownership-aware lifecycle safeguard.

After upgrading, review and Trust the modified UserPromptSubmit Hook in Settings → Hooks, fully restart Codex Desktop, and start a new task.

Multica Agent Sync 1.0.1

Choose a tag to compare

@QuanFeel QuanFeel released this 20 Jul 06:48
7fe6928

Patch release fixing operational Skill data discovery.

Fixed

  • status, doctor, and cleanup now find Codex-managed plugin data when their Skill terminal does not provide PLUGIN_DATA.
  • The runtime no longer falls back to the obsolete ~/.multica/plugin-data location.
  • Hook and command entry points share the same safe path resolver.

Install latest stable

# Step 1 of 2: register the marketplace at the latest stable version.
codex plugin marketplace add zhongwangquan/multica-agent-sync

# Step 2 of 2: install and enable the plugin.
codex plugin add multica-codex-sync@multica-agent-sync

Optional exact version

# Optional step 1 of 2: pin the marketplace to v1.0.1.
codex plugin marketplace add zhongwangquan/multica-agent-sync --ref v1.0.1

# Step 2 of 2: install and enable that exact version.
codex plugin add multica-codex-sync@multica-agent-sync

Validated with 27 functional and security tests, Ruff, the official Codex plugin validator, isolated installs, and a live read-only status check against an existing tracker.

Multica Agent Sync 1.0.0

Choose a tag to compare

@QuanFeel QuanFeel released this 20 Jul 06:07
6fae7fe

First public release of the Codex-only Multica Agent Sync plugin.

Highlights:

  • Install from the public Codex Git marketplace, including optional exact-tag installs.
  • Bind one Codex Desktop task to one Multica issue with /multica.
  • Private plugin-owned runtime data and ownership-aware cleanup.
  • 26 functional and security tests, release validation, and macOS CI.
  • English and Chinese documentation.

Normal installation — both steps are required the first time:

# Step 1 of 2: register the marketplace at the latest stable version.
codex plugin marketplace add zhongwangquan/multica-agent-sync

# Step 2 of 2: install and enable the plugin from that marketplace.
codex plugin add multica-codex-sync@multica-agent-sync

Optional exact-version installation or rollback:

# Optional step 1 of 2: register the marketplace at v1.0.0.
codex plugin marketplace add zhongwangquan/multica-agent-sync --ref v1.0.0

# Step 2 of 2: install and enable that exact plugin version.
codex plugin add multica-codex-sync@multica-agent-sync

No extra package is required. Codex installs the plugin source from Git; GitHub provides source .zip and .tar.gz archives automatically.