bug(plan): `Plan.can_revert_to()` incorrectly rejects `STRATEGIZE → STRATEGIZE` reversion — errored Strategize plans cannot be restarted #2380

New issue

Open

opened 2026-04-03 17:26:49 +00:00 by freemo · 0 comments

freemo commented

2026-04-03 17:26:49 +00:00

Owner

Metadata

Branch: fix/plan-can-revert-to-strategize-self
Commit Message: fix(plan): allow STRATEGIZE→STRATEGIZE reversion in can_revert_to() for errored plans
Milestone: v3.5.0
Parent Epic: #368

Background and Context

Plan.can_revert_to() in src/cleveragents/domain/models/core/plan.py delegates to can_transition(self.phase, phase) to validate whether a reversion is allowed. The VALID_PHASE_TRANSITIONS map is:

VALID_PHASE_TRANSITIONS: dict[PlanPhase, list[PlanPhase]] = {
    PlanPhase.ACTION: [PlanPhase.STRATEGIZE],      # plan use
    PlanPhase.STRATEGIZE: [PlanPhase.EXECUTE],     # execute command
    PlanPhase.EXECUTE: [PlanPhase.APPLY, PlanPhase.STRATEGIZE],  # apply or revert
    PlanPhase.APPLY: [PlanPhase.STRATEGIZE],       # revert via constrained
}

This means:

EXECUTE → STRATEGIZE is allowed ✅ (Execute can revert to Strategize)
APPLY → STRATEGIZE is allowed ✅ (constrained Apply can revert to Strategize)
STRATEGIZE → STRATEGIZE is NOT in the map ❌ (missing)

The problem: after a plan is reverted from Execute back to Strategize, it is now in the Strategize phase. If the Strategize phase then fails (ERRORED state), the can_revert_to(PlanPhase.STRATEGIZE) call returns False because STRATEGIZE → STRATEGIZE is not a valid transition. This means a plan that has been reverted to Strategize and then errored in Strategize cannot be re-run in Strategize — it is permanently stuck.

The spec states: "ERRORED and CONSTRAINED plans can still be reverted to an earlier phase via revert_plan or resumed through PlanResumeService." A plan in STRATEGIZE/ERRORED state should be able to revert to STRATEGIZE (i.e., restart Strategize).

Reproduction Steps

Create and use an action to create a plan (now in STRATEGIZE/QUEUED)
Start strategize: plan transitions to STRATEGIZE/PROCESSING
Fail strategize: plan transitions to STRATEGIZE/ERRORED
Try to revert to Strategize: plan.can_revert_to(PlanPhase.STRATEGIZE) returns False
The plan is stuck — cannot be restarted in Strategize

Alternatively, the same bug is triggered via the revert path:

Plan progresses to EXECUTE/ERRORED
Revert to Strategize: plan is now in STRATEGIZE/QUEUED
Execute Strategize: plan transitions to STRATEGIZE/PROCESSING
Strategize fails: plan transitions to STRATEGIZE/ERRORED
can_revert_to(PlanPhase.STRATEGIZE) returns False — plan is stuck

Current Behavior

Plan.can_revert_to(PlanPhase.STRATEGIZE) returns False when the plan is already in STRATEGIZE phase (regardless of processing state), because VALID_PHASE_TRANSITIONS[STRATEGIZE] only contains [EXECUTE], not [STRATEGIZE].

Code locations:

src/cleveragents/domain/models/core/plan.py — VALID_PHASE_TRANSITIONS dict (~line 1185)
src/cleveragents/domain/models/core/plan.py — can_revert_to() method (~line 1060)

Expected Behavior

A plan in STRATEGIZE/ERRORED or STRATEGIZE/CONSTRAINED state should be able to call can_revert_to(PlanPhase.STRATEGIZE) and receive True, allowing the plan to be restarted in the Strategize phase. The fix should be scoped to can_revert_to() (not can_transition()), since STRATEGIZE → STRATEGIZE is not a valid forward transition — only a valid self-reversion when in a recoverable terminal state.

Subtasks

Investigate can_revert_to() and can_transition() in plan.py to understand the full delegation chain
Fix can_revert_to() to allow STRATEGIZE → STRATEGIZE when the plan's processing state is ERRORED or CONSTRAINED
Ensure can_transition() is not modified (forward transitions must remain unchanged)
Add unit tests for can_revert_to(PlanPhase.STRATEGIZE) when plan is in STRATEGIZE/ERRORED state
Add unit tests for can_revert_to(PlanPhase.STRATEGIZE) when plan is in STRATEGIZE/CONSTRAINED state
Add unit tests confirming can_revert_to(PlanPhase.STRATEGIZE) still returns False for STRATEGIZE/PROCESSING and STRATEGIZE/QUEUED (non-terminal states should not be revertable)
Tests (Behave): Add BDD scenario for reverting an errored Strategize plan back to Strategize
Verify coverage >= 97% via nox -s coverage_report
Run nox (all default sessions), fix any errors

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%.

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-new-issue-creator

## Metadata - **Branch**: `fix/plan-can-revert-to-strategize-self` - **Commit Message**: `fix(plan): allow STRATEGIZE→STRATEGIZE reversion in can_revert_to() for errored plans` - **Milestone**: v3.5.0 - **Parent Epic**: #368 ## Background and Context `Plan.can_revert_to()` in `src/cleveragents/domain/models/core/plan.py` delegates to `can_transition(self.phase, phase)` to validate whether a reversion is allowed. The `VALID_PHASE_TRANSITIONS` map is: ```python VALID_PHASE_TRANSITIONS: dict[PlanPhase, list[PlanPhase]] = { PlanPhase.ACTION: [PlanPhase.STRATEGIZE], # plan use PlanPhase.STRATEGIZE: [PlanPhase.EXECUTE], # execute command PlanPhase.EXECUTE: [PlanPhase.APPLY, PlanPhase.STRATEGIZE], # apply or revert PlanPhase.APPLY: [PlanPhase.STRATEGIZE], # revert via constrained } ``` This means: - `EXECUTE → STRATEGIZE` is allowed ✅ (Execute can revert to Strategize) - `APPLY → STRATEGIZE` is allowed ✅ (constrained Apply can revert to Strategize) - `STRATEGIZE → STRATEGIZE` is **NOT** in the map ❌ (missing) **The problem**: after a plan is reverted from Execute back to Strategize, it is now in the Strategize phase. If the Strategize phase then fails (`ERRORED` state), the `can_revert_to(PlanPhase.STRATEGIZE)` call returns `False` because `STRATEGIZE → STRATEGIZE` is not a valid transition. This means a plan that has been reverted to Strategize and then errored in Strategize cannot be re-run in Strategize — it is permanently stuck. The spec states: *"ERRORED and CONSTRAINED plans can still be reverted to an earlier phase via revert_plan or resumed through PlanResumeService."* A plan in `STRATEGIZE/ERRORED` state should be able to revert to `STRATEGIZE` (i.e., restart Strategize). ## Reproduction Steps 1. Create and use an action to create a plan (now in `STRATEGIZE/QUEUED`) 2. Start strategize: plan transitions to `STRATEGIZE/PROCESSING` 3. Fail strategize: plan transitions to `STRATEGIZE/ERRORED` 4. Try to revert to Strategize: `plan.can_revert_to(PlanPhase.STRATEGIZE)` returns `False` 5. The plan is stuck — cannot be restarted in Strategize Alternatively, the same bug is triggered via the revert path: 1. Plan progresses to `EXECUTE/ERRORED` 2. Revert to Strategize: plan is now in `STRATEGIZE/QUEUED` 3. Execute Strategize: plan transitions to `STRATEGIZE/PROCESSING` 4. Strategize fails: plan transitions to `STRATEGIZE/ERRORED` 5. `can_revert_to(PlanPhase.STRATEGIZE)` returns `False` — plan is stuck ## Current Behavior `Plan.can_revert_to(PlanPhase.STRATEGIZE)` returns `False` when the plan is already in `STRATEGIZE` phase (regardless of processing state), because `VALID_PHASE_TRANSITIONS[STRATEGIZE]` only contains `[EXECUTE]`, not `[STRATEGIZE]`. **Code locations**: - `src/cleveragents/domain/models/core/plan.py` — `VALID_PHASE_TRANSITIONS` dict (~line 1185) - `src/cleveragents/domain/models/core/plan.py` — `can_revert_to()` method (~line 1060) ## Expected Behavior A plan in `STRATEGIZE/ERRORED` or `STRATEGIZE/CONSTRAINED` state should be able to call `can_revert_to(PlanPhase.STRATEGIZE)` and receive `True`, allowing the plan to be restarted in the Strategize phase. The fix should be scoped to `can_revert_to()` (not `can_transition()`), since `STRATEGIZE → STRATEGIZE` is not a valid forward transition — only a valid self-reversion when in a recoverable terminal state. ## Subtasks - [ ] Investigate `can_revert_to()` and `can_transition()` in `plan.py` to understand the full delegation chain - [ ] Fix `can_revert_to()` to allow `STRATEGIZE → STRATEGIZE` when the plan's processing state is `ERRORED` or `CONSTRAINED` - [ ] Ensure `can_transition()` is **not** modified (forward transitions must remain unchanged) - [ ] Add unit tests for `can_revert_to(PlanPhase.STRATEGIZE)` when plan is in `STRATEGIZE/ERRORED` state - [ ] Add unit tests for `can_revert_to(PlanPhase.STRATEGIZE)` when plan is in `STRATEGIZE/CONSTRAINED` state - [ ] Add unit tests confirming `can_revert_to(PlanPhase.STRATEGIZE)` still returns `False` for `STRATEGIZE/PROCESSING` and `STRATEGIZE/QUEUED` (non-terminal states should not be revertable) - [ ] Tests (Behave): Add BDD scenario for reverting an errored Strategize plan back to Strategize - [ ] Verify coverage >= 97% via `nox -s coverage_report` - [ ] Run `nox` (all default sessions), fix any errors ## 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%. --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-new-issue-creator

freemo added this to the v3.5.0 milestone

2026-04-03 17:26:55 +00:00

freemo added the

labels

2026-04-03 17:26:59 +00:00

freemo added a new dependency