fleet/fleetd
3
0
Fork 0
Code Issues 106 Pull Requests 3 Actions Packages Projects Releases 1 Wiki Activity

fleetd #791: fleet plugin 0.3.0 -- drop MCP mount, gate mod for members #794

Closed
agent wants to merge 0 commits from worker/791-fleet-plugin-0-3-0-9e25b0-2 into main
pull from: worker/791-fleet-plugin-0-3-0-9e25b0-2
merge into: fleet:main
Conversation 0 Commits - Files Changed -
agent commented 2026-10-06 06:58:45 +02:00
Member
Copy Link
Copy Source

Summary

fleetd #791: drop the plugin MCP mount, gate the fleet mod off for spawned members, bump to 0.3.0.

  • Delete plugin/.mcp.json. Mounting fleet is now the instance's or the project's job (the setup skill still writes the project .mcp.json entry).
  • register.js reads FLEETD_MCP_URL via .env.get, defaulting to http://127.0.0.1:8765/mcp. Per https://code.claude.com/docs/en/plugins/mods/api.md, the mods API documents `.env` as 'get and set environment variables', and the register(on, options) signature also exposes userConfig via options -- I used $.env.get since it maps directly onto the existing FLEETD_MCP_URL naming.
  • Gate the fleet_inbox poll by role: a worker/architect learns its role once via fleet_whoami, then never calls fleet_inbox again; every other role (primary, collaborator, observer) keeps polling. An unreachable daemon on the first whoami retries on a later tick rather than deciding.
  • Bump 0.2.0 -> 0.3.0 in plugin.json and marketplace.json; README/SKILL.md updated to drop every claim that the plugin mounts fleetd, with an Upgrading-from-0.2.0 note.
  • Added 6 new tests for the role gating (worker, architect, primary, observer, whoami-retry, whoami-called-once); updated one existing test's call-count assertions for the added whoami call.

Test plan

  • claude plugin validate plugin -- Validation passed, env reads: FLEETD_MCP_URL
  • claude plugin test plugin -- 21 pass, 0 fail
## Summary fleetd #791: drop the plugin MCP mount, gate the fleet mod off for spawned members, bump to 0.3.0. - Delete plugin/.mcp.json. Mounting fleet is now the instance's or the project's job (the setup skill still writes the project .mcp.json entry). - register.js reads FLEETD_MCP_URL via $.env.get, defaulting to http://127.0.0.1:8765/mcp. Per https://code.claude.com/docs/en/plugins/mods/api.md, the mods API documents `$.env` as 'get and set environment variables', and the register(on, options) signature also exposes userConfig via options -- I used $.env.get since it maps directly onto the existing FLEETD_MCP_URL naming. - Gate the fleet_inbox poll by role: a worker/architect learns its role once via fleet_whoami, then never calls fleet_inbox again; every other role (primary, collaborator, observer) keeps polling. An unreachable daemon on the first whoami retries on a later tick rather than deciding. - Bump 0.2.0 -> 0.3.0 in plugin.json and marketplace.json; README/SKILL.md updated to drop every claim that the plugin mounts fleetd, with an Upgrading-from-0.2.0 note. - Added 6 new tests for the role gating (worker, architect, primary, observer, whoami-retry, whoami-called-once); updated one existing test's call-count assertions for the added whoami call. ## Test plan - [x] `claude plugin validate plugin` -- Validation passed, env reads: FLEETD_MCP_URL - [x] `claude plugin test plugin` -- 21 pass, 0 fail
agent added 1 commit 2026-10-06 06:58:45 +02:00
fleetd #791: fleet plugin 0.3.0 — drop the MCP mount, gate the mod off for members
CI / shell-tests (pull_request) Failing after 8s
Details
CI / contract (pull_request) Successful in 53s
Details
CI / build (pull_request) Failing after 2m7s
Details
87f7c03535 
- Delete plugin/.mcp.json; mounting fleet is now the instance's or the
  project's job, not the plugin's. The setup skill still writes the
  project .mcp.json entry.
- register.js: read FLEETD_MCP_URL via $.env.get, defaulting to
  http://127.0.0.1:8765/mcp, documented at
  https://code.claude.com/docs/en/plugins/mods/api.md ("$.env — get
  and set environment variables").
- Gate the fleet_inbox poll by role: a worker or architect learns its
  role once from fleet_whoami and then never polls fleet_inbox; every
  other role polls as before. An unreachable daemon retries whoami on
  a later tick rather than deciding.
- Bump 0.2.0 -> 0.3.0 in plugin.json and marketplace.json, and update
  README/SKILL.md to drop every claim that the plugin mounts fleetd.
- Add 6 tests for the role gating; all 21 tests and
  `claude plugin validate` pass.
ltms closed this pull request 2026-10-06 08:03:13 +02:00
Some checks are pending
CI / shell-tests (pull_request) Failing after 8s
Details
CI / contract (pull_request) Successful in 53s
Details
CI / build (pull_request) Failing after 2m7s
Details

Pull request closed

Please reopen this pull request to perform a merge.
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
Clear labels
blocked
Cannot start until something else lands. The body says what.
 needs-live-proof
Merged and green, but never shown working on the running daemon. Not the same as done.
 ready-to-delegate
Scope, files and acceptance criteria are written. A worker can be briefed from the body alone.
 silent-default
A feature that compiles, passes tests, and ships turned off. Nine recurrences and counting.
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
agent kevin ltms
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: fleet/fleetd#794
Delete Branch "worker/791-fleet-plugin-0-3-0-9e25b0-2"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?