cleveragents/cleveragents-core

Fork 3

[BUG] agents plan use leaves plan in strategize/queued — Strategize phase never auto-starts #9188

New issue

Open

opened 2026-04-14 09:48:28 +00:00 by hurui200320 · 2 comments

hurui200320 commented

2026-04-14 09:48:28 +00:00

Member

Metadata

Branch: fix/plan-use-strategize-never-auto-starts
Commit Message: fix(plan): auto-start Strategize phase after agents plan use instead of leaving plan in queued state
Milestone: v3.2.0
Parent Epic: TBD — requires manual linking

Description

Background

agents plan use is the entry point for the four-phase plan lifecycle (Action → Strategize → Execute → Apply). Per the spec (agents plan use section): "Apply an action to one or more projects and start the Strategize phase." The Strategize phase is intended to be a read-only, inspectable phase where the system builds a decision tree before any execution begins — this separation is a core design guarantee.

Current Behavior

After running:

agents plan use <action> <project> --automation-profile trusted

the plan is created but its processing_state is permanently queued. The Strategize phase never starts automatically. The CLI output itself confirms this:

Plan is now in Strategize phase (queued). Run 'agents plan execute <id>' when ready.

Polling agents plan status <PLAN_ID> repeatedly confirms the state stays strategize/queued indefinitely — it never transitions to running or completed on its own.

Additional observation: Running agents plan execute <PLAN_ID> drives both the Strategize phase AND the Execute phase in a single call — both phases complete within that one command. This indicates either the phase model has been collapsed, or the async worker responsible for auto-starting Strategize is not running.

Actual flow observed:

agents plan use ... → plan created, state: strategize/queued (no auto-start)
agents plan status <ID> → phase: strategize, state: queued (indefinitely)
agents plan tree <ID> → "No decisions found"
agents plan execute <ID> → drives both Strategize and Execute, then execute/complete

Reproduction steps:

agents resource add git-checkout local/rune-project --path /workspace/rune
agents project create --resource local/rune-project local/rune-encoder
agents action create --config /workspace/build-rune-encoder.yaml
agents plan use local/build-rune-encoder local/rune-encoder --automation-profile trusted
Wait 60+ seconds, then agents plan status <PLAN_ID> → state remains strategize/queued
agents plan tree <PLAN_ID> → "No decisions found"

Expected Behavior

Per spec §agents plan use: "Apply an action to one or more projects and start the Strategize phase."

The Strategize phase should begin automatically after plan use, with agents plan status showing phase: strategize, state: running (and eventually completed) without any additional commands. Only then should agents plan execute be needed to advance to the Execute phase.

Expected flow per spec:

agents plan use ... → plan created, Strategize starts automatically
agents plan status <ID> → phase: strategize, state: running or completed
agents plan tree <ID> → decision tree populated
agents plan execute <ID> → Execute phase starts

Acceptance Criteria

After agents plan use, the plan's Strategize phase starts automatically (state transitions from queued → running → completed without any additional commands)
agents plan status <ID> shows phase: strategize, state: running (or completed) after plan use completes
agents plan tree <ID> returns a populated decision tree after Strategize completes
agents plan execute <ID> only drives the Execute phase (not Strategize), consistent with the four-phase model
The CLI output of agents plan use does not instruct the user to run agents plan execute to start Strategize

Supporting Information

Related issue (distinct bug): #5426 — covers plans stuck in Strategize/PROCESSING being invisible to auto-discovery; that is a filtering bug, not the Strategize auto-start failure described here.
Spec reference: docs/specification.md §agents plan use — "Apply an action to one or more projects and start the Strategize phase."

Subtasks

Identify root cause: determine whether the async worker for Strategize auto-start is missing, disabled, or the phase model has been collapsed in plan use implementation
Restore or implement the Strategize auto-start trigger in the plan use command handler / service layer
Ensure agents plan execute only drives the Execute phase (not Strategize) after the fix
Update CLI output of agents plan use to reflect correct next-step messaging (remove instruction to run plan execute to start Strategize)
Tests: add/update BDD scenarios covering the plan use → Strategize auto-start → plan tree populated flow
Tests: add/update unit tests for the Strategize auto-start trigger
Run nox (all default sessions), fix any errors
Verify coverage >= 97% via nox -s coverage_report

Definition of Done

This issue is complete when:

All subtasks above are completed and checked off.
A Git commit is created where the first line of the commit message matches the Commit Message in Metadata exactly, followed by a blank line, then additional lines providing relevant details about the implementation.
The commit is pushed to the remote on the branch matching the Branch in Metadata exactly.
The commit is submitted as a pull request to master, reviewed, and merged before this issue is marked done.
All nox stages pass.
Coverage >= 97%.

## Metadata - **Branch**: `fix/plan-use-strategize-never-auto-starts` - **Commit Message**: `fix(plan): auto-start Strategize phase after agents plan use instead of leaving plan in queued state` - **Milestone**: v3.2.0 - **Parent Epic**: TBD — requires manual linking ## Description ### Background `agents plan use` is the entry point for the four-phase plan lifecycle (Action → Strategize → Execute → Apply). Per the spec (`agents plan use` section): *"Apply an action to one or more projects and start the Strategize phase."* The Strategize phase is intended to be a read-only, inspectable phase where the system builds a decision tree before any execution begins — this separation is a core design guarantee. ### Current Behavior After running: ``` agents plan use <action> <project> --automation-profile trusted ``` the plan is created but its `processing_state` is permanently `queued`. The Strategize phase never starts automatically. The CLI output itself confirms this: ``` Plan is now in Strategize phase (queued). Run 'agents plan execute <id>' when ready. ``` Polling `agents plan status <PLAN_ID>` repeatedly confirms the state stays `strategize/queued` indefinitely — it never transitions to `running` or `completed` on its own. **Additional observation:** Running `agents plan execute <PLAN_ID>` drives **both** the Strategize phase AND the Execute phase in a single call — both phases complete within that one command. This indicates either the phase model has been collapsed, or the async worker responsible for auto-starting Strategize is not running. **Actual flow observed:** 1. `agents plan use ...` → plan created, state: `strategize/queued` (no auto-start) 2. `agents plan status <ID>` → `phase: strategize, state: queued` (indefinitely) 3. `agents plan tree <ID>` → "No decisions found" 4. `agents plan execute <ID>` → drives **both** Strategize and Execute, then `execute/complete` **Reproduction steps:** 1. `agents resource add git-checkout local/rune-project --path /workspace/rune` 2. `agents project create --resource local/rune-project local/rune-encoder` 3. `agents action create --config /workspace/build-rune-encoder.yaml` 4. `agents plan use local/build-rune-encoder local/rune-encoder --automation-profile trusted` 5. Wait 60+ seconds, then `agents plan status <PLAN_ID>` → state remains `strategize/queued` 6. `agents plan tree <PLAN_ID>` → "No decisions found" ### Expected Behavior Per spec §`agents plan use`: *"Apply an action to one or more projects and start the Strategize phase."* The Strategize phase should begin automatically after `plan use`, with `agents plan status` showing `phase: strategize, state: running` (and eventually `completed`) without any additional commands. Only then should `agents plan execute` be needed to advance to the Execute phase. **Expected flow per spec:** 1. `agents plan use ...` → plan created, Strategize starts automatically 2. `agents plan status <ID>` → `phase: strategize, state: running` or `completed` 3. `agents plan tree <ID>` → decision tree populated 4. `agents plan execute <ID>` → Execute phase starts ### Acceptance Criteria - [ ] After `agents plan use`, the plan's Strategize phase starts automatically (state transitions from `queued` → `running` → `completed` without any additional commands) - [ ] `agents plan status <ID>` shows `phase: strategize, state: running` (or `completed`) after `plan use` completes - [ ] `agents plan tree <ID>` returns a populated decision tree after Strategize completes - [ ] `agents plan execute <ID>` only drives the Execute phase (not Strategize), consistent with the four-phase model - [ ] The CLI output of `agents plan use` does not instruct the user to run `agents plan execute` to start Strategize ### Supporting Information - **Related issue (distinct bug):** #5426 — covers plans stuck in `Strategize/PROCESSING` being invisible to auto-discovery; that is a filtering bug, not the Strategize auto-start failure described here. - **Spec reference:** `docs/specification.md` §`agents plan use` — *"Apply an action to one or more projects and start the Strategize phase."* ## Subtasks - [ ] Identify root cause: determine whether the async worker for Strategize auto-start is missing, disabled, or the phase model has been collapsed in `plan use` implementation - [ ] Restore or implement the Strategize auto-start trigger in the `plan use` command handler / service layer - [ ] Ensure `agents plan execute` only drives the Execute phase (not Strategize) after the fix - [ ] Update CLI output of `agents plan use` to reflect correct next-step messaging (remove instruction to run `plan execute` to start Strategize) - [ ] Tests: add/update BDD scenarios covering the `plan use` → Strategize auto-start → `plan tree` populated flow - [ ] Tests: add/update unit tests for the Strategize auto-start trigger - [ ] Run `nox` (all default sessions), fix any errors - [ ] Verify coverage >= 97% via `nox -s coverage_report` ## Definition of Done This issue is complete when: - All subtasks above are completed and checked off. - A Git commit is created where the **first line** of the commit message matches the Commit Message in Metadata exactly, followed by a blank line, then additional lines providing relevant details about the implementation. - The commit is pushed to the remote on the branch matching the **Branch** in Metadata exactly. - The commit is submitted as a **pull request** to `master`, reviewed, and **merged** before this issue is marked done. - All nox stages pass. - Coverage >= 97%.

hurui200320 added the

labels

2026-04-14 09:48:43 +00:00

hurui200320 added this to the v3.2.0 milestone

2026-04-14 09:48:48 +00:00

HAL9000 commented

2026-04-14 10:27:28 +00:00

Owner

@hurui200320 Thank you for this bug report.

Acknowledgment — Issue #9188

This issue has been received. The reported behavior — agents plan use leaving the plan in strategize/queued state without auto-starting the Strategize phase — has been logged.

Current status: State/Unverified — this will be reviewed by the project owner for verification and prioritization.

Next steps: The project owner will verify this issue. Given the Priority/High label, this should be addressed promptly once verified.

Automated by CleverAgents Bot
Supervisor: Human Liaison | Agent: human-liaison-pool-supervisor
Worker: [AUTO-HUMAN-7]

@hurui200320 Thank you for this bug report. **Acknowledgment — Issue #9188** This issue has been received. The reported behavior — `agents plan use` leaving the plan in strategize/queued state without auto-starting the Strategize phase — has been logged. **Current status**: State/Unverified — this will be reviewed by the project owner for verification and prioritization. **Next steps**: The project owner will verify this issue. Given the Priority/High label, this should be addressed promptly once verified. --- **Automated by CleverAgents Bot** Supervisor: Human Liaison | Agent: human-liaison-pool-supervisor Worker: [AUTO-HUMAN-7]

HAL9000 added

and removed

labels

2026-04-14 12:49:06 +00:00

HAL9000 commented

2026-04-14 12:49:06 +00:00

Owner

✅ Triage: Verified [AUTO-OWNR-1]

This is a critical bug in the core plan lifecycle. Per the spec, agents plan use should start the Strategize phase automatically — the current behavior leaves plans permanently in strategize/queued state, breaking the four-phase lifecycle model. The additional observation that agents plan execute drives both Strategize and Execute phases in a single call confirms the async worker for Strategize auto-start is missing or disabled.

Assigning to v3.2.0 as this is a core plan lifecycle bug that must be resolved for M3's decision tree and invariant features to function correctly. Priority is High — this breaks the intended workflow.

MoSCoW: Must Have — the four-phase plan lifecycle is a core architectural guarantee. The milestone cannot be considered complete without this working correctly.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: project-owner-pool-supervisor

✅ **Triage: Verified** [AUTO-OWNR-1] This is a critical bug in the core plan lifecycle. Per the spec, `agents plan use` should start the Strategize phase automatically — the current behavior leaves plans permanently in `strategize/queued` state, breaking the four-phase lifecycle model. The additional observation that `agents plan execute` drives both Strategize and Execute phases in a single call confirms the async worker for Strategize auto-start is missing or disabled. Assigning to **v3.2.0** as this is a core plan lifecycle bug that must be resolved for M3's decision tree and invariant features to function correctly. Priority is **High** — this breaks the intended workflow. MoSCoW: **Must Have** — the four-phase plan lifecycle is a core architectural guarantee. The milestone cannot be considered complete without this working correctly. --- **Automated by CleverAgents Bot** Supervisor: Project Owner | Agent: project-owner-pool-supervisor

HAL9000 referenced this issue

2026-04-14 12:51:00 +00:00

[AUTO-PROJ-OWN] Status: Project Owner Pool Supervisor (Cycle 1) #9249

HAL9000 referenced this issue

2026-04-14 12:56:58 +00:00

[AUTO-PROJ-OWN] Status: Project Owner Pool Supervisor (Cycle 1) #9249

HAL9000 referenced this issue

2026-04-14 14:44:35 +00:00

[AUTO-IMP-POOL] Status: Implementation Pool Supervisor (Cycle 5) #9318