Skill: Stale Flow Cleanup

Purpose

Weekly maintenance pass to abandon teaching flows that have been inactive for 30+ days and clean
up their associated pending spaced repetition schedules. Prevents orphaned flows from cluttering
the active state and schedules from firing for topics the user has effectively stopped studying.

When to Use

Use this skill when:

• The weekly-stale-flow-check scheduled task fires (cron: 0 4 * * 1, Mondays at 04:00)

Staleness Criteria

A teaching flow is stale when all of the following are true:

• status is active (i.e., not completed or abandoned)
• last_session_at is more than 30 days ago (or null and created_at is more than 30 days ago)

Cleanup Protocol

Step 1: List Active Flows

Call teaching_flow_list(status="active") to retrieve all active flows.

The response includes, per flow: mind_map_id, mind_map_title, status, created_at,
last_session_at.

If no active flows are returned, exit silently — no notification needed for a maintenance no-op.

Step 2: Filter for Stale Flows

From the active flows, identify stale flows:

from datetime import datetime, timezone, timedelta

STALE_THRESHOLD_DAYS = 30
now = datetime.now(timezone.utc)

stale_flows = [
    flow for flow in active_flows
    if (flow["last_session_at"] is not None
        and (now - datetime.fromisoformat(flow["last_session_at"])).days > STALE_THRESHOLD_DAYS)
    or (flow["last_session_at"] is None
        and (now - datetime.fromisoformat(flow["created_at"])).days > STALE_THRESHOLD_DAYS)
]

If no stale flows are found, exit without taking further action.

Step 3: Abandon Each Stale Flow

For each stale flow, in sequence:

1. Call teaching_flow_abandon(mind_map_id=<mind_map_id>) to transition the flow status from
   active to abandoned.

2. Call spaced_repetition_schedule_cleanup(mind_map_id=<mind_map_id>) to remove all pending
   review schedules associated with this mind map.

3. Call memory_store_fact() to record the abandonment:

   memory_store_fact(
       subject=<mind_map_title>,
       predicate="study_pattern",
       content=f"Teaching flow abandoned after 30+ days of inactivity. "
               f"Last active: {flow['last_session_at'] or 'never'}.",
       permanence="volatile",
       importance=4.0,
       tags=[<topic_tag_derived_from_title>, "paused", "stale-flow-cleanup"]
   )

Step 4: Notify the User

After processing all stale flows, send a summary notification:

notify(
    channel="telegram",
    intent="send",
    message=f"Weekly cleanup: {len(stale_flows)} stale learning flow(s) archived after 30+ days "
            f"of inactivity — {', '.join(f['mind_map_title'] for f in stale_flows)}. "
            f"Your progress is preserved. Say 'resume [topic]' anytime to pick up where you left off.",
)

If no stale flows were found, skip this notification (no news is good news for a maintenance task).

Exit Criteria

• teaching_flow_list(status="active") was called to retrieve all active flows
• All flows inactive for 30+ days have been identified
• teaching_flow_abandon() called for each stale flow
• spaced_repetition_schedule_cleanup() called for each stale flow to remove pending reviews
• A memory_store_fact() with predicate="study_pattern" recorded for each abandoned flow
• User notified of cleanup summary (only if at least one flow was abandoned)
• Session exits without teaching, reviewing, or modifying non-stale flows