Workflow Transition

Execution Context: This skill runs in a subagent session, delegated by the PAW orchestrator when processing a paw-transition TODO. Return structured output—do not make orchestration decisions beyond the transition.

Purpose

Gate between workflow activities. Ensures:

• Session policy is respected at stage boundaries
• Correct branch state before implementation
• Required artifacts exist before proceeding
• Next activity is correctly identified and queued

Procedure

Execute these steps in order. Do not skip steps.

Step 1: Identify Current State

Read WorkflowContext.md to determine:

• Work ID and target branch
• Session Policy (per-stage | continuous)
• Review Strategy (prs | local)
• Review Policy (every-stage | milestones | planning-only | final-pr-only)
  • If missing, check for legacy Handoff Mode: field and map: manual→every-stage, semi-auto→milestones, auto→final-pr-only
  • Also map legacy Review Policy values: always→every-stage, never→final-pr-only
  • If neither present, default to milestones
• Final Agent Review (enabled | disabled)
  • If missing, default to enabled

Identify last completed activity from TODOs or artifacts.

Step 2: Determine Next Activity

Use the Mandatory Transitions table:

Skippable = NO: Add activity TODO and execute immediately after transition completes.

Step 2.5: Candidate Promotion Check

When next activity would be paw-pr (all planned phases complete), check for phase candidates:

1. Read ImplementationPlan.md ## Phase Candidates section
2. Count unresolved candidates: - [ ] items WITHOUT terminal tags ([skipped], [deferred], [not feasible])
3. If unresolved candidates exist: set promotion_pending = true and extract candidate descriptions
4. Otherwise: set promotion_pending = false

If promotion_pending = true, return candidates in structured output. PAW orchestrator handles user interaction.

Step 3: Check Stage Boundary and Milestone Pause

Stage boundaries occur when moving between these stages:

• spec-review passes → code-research
• plan-review passes → implement (Phase 1)
• paw-planning-docs-review complete → implement (Phase 1)
• phase N complete → phase N+1
• all phases complete → paw-final-review (if enabled) or paw-pr (if disabled)
• paw-final-review complete → paw-pr
• paw-pr complete → workflow complete

Stage-to-milestone mapping (determines which milestone is reached at each boundary):

Determine pause_at_milestone:

• If Review Policy ∈ {every-stage, milestones}: pause at ALL milestones
• If Review Policy = planning-only:
  • Spec.md, ImplementationPlan.md, Planning Documents Review complete, Final PR: pause_at_milestone = true
  • Phase completion (including last phase): pause_at_milestone = false
• If Review Policy = final-pr-only:
  • Final PR: pause_at_milestone = true
  • All other milestones: pause_at_milestone = false

Determine session_action:

• If crossing a stage boundary AND Session Policy = per-stage: set session_action = new_session
• Otherwise: set session_action = continue

If session_action = new_session, set inline_instruction to: next activity and phase (e.g., "Phase 2: Tool Enhancement")

Continue to Step 4 (preflight still needed for inline_instruction context).

Step 4: Preflight Checks

Before the next activity can start, verify:

For paw-implement:

• On correct branch per Review Strategy
  • prs: phase branch (e.g., <target>_phase1)
  • local: target branch
• ImplementationPlan.md exists and has the target phase

For paw-code-research:

• Spec.md exists (unless minimal mode)

For paw-planning-docs-review:

• Spec.md exists (unless minimal mode)
• ImplementationPlan.md exists

For paw-final-review:

• All implementation phases complete
• On target branch (local strategy) or phase branches merged (prs strategy)

For paw-pr:

• All phases complete
• paw-final-review complete (if enabled) or skipped (if disabled)
• On target branch or ready to merge

Artifact Tracking Check (for all activities):

• Check if .paw/work/<work-id>/.gitignore exists
• If exists with * pattern: artifact_tracking = disabled
• Otherwise: artifact_tracking = enabled

If any check fails, report blocker and stop.

Step 5: Queue Next Activity

Add TODO for next activity:

• [ ] <activity-name> (<context>)
• [ ] paw-transition

Completion

After completing all steps, return structured output:

TRANSITION RESULT:
- session_action: [continue | new_session]
- pause_at_milestone: [true | false]
- next_activity: [activity name and context]
- artifact_tracking: [enabled | disabled]
- preflight: [passed | blocked: <reason>]
- work_id: [current work ID]
- inline_instruction: [for new_session only: resume hint]
- promotion_pending: [true | false] (only when next would be paw-pr)
- candidates: [list of unresolved candidate descriptions] (only if promotion_pending)

If pause_at_milestone = true: The PAW agent must PAUSE and wait for user confirmation before proceeding.

If session_action = new_session: The PAW agent must call paw_new_session with the provided work_id and inline_instruction.

If preflight = blocked: The PAW agent must report the blocker to the user.

Mark the paw-transition TODO complete after returning this output.

Candidate Resolution Markers

Terminal markers used in Step 2.5 to identify resolved candidates:

• - [x] [promoted] <desc> — Elaborated into a full phase
• - [x] [skipped] <desc> — User chose not to pursue
• - [x] [deferred] <desc> — Future work outside current workflow
• - [x] [not feasible] <desc> — Code research revealed infeasibility

Unresolved: - [ ] items without terminal tags. Empty section or all-resolved → promotion_pending = false.

Guardrails

• Do NOT skip the stage boundary check (Step 3)
• Do NOT return session_action = continue if boundary + per-stage policy
• Do NOT return preflight = passed if checks actually failed
• Do NOT call paw_new_session directly—return the decision for PAW agent to act on