Reopening Tasks

Reopen completed tasks properly with time tracking and requirement documentation.

Announce at start: "I'm using the knowns.task.reopen skill to reopen task [ID]."

Core principle: DOCUMENT WHY THE TASK IS REOPENED.

The Process

Step 1: View Current Task State

json

mcp__knowns__get_task({ "taskId": "$ARGUMENTS" })

Verify:

•Task is currently done
•Understand what was implemented
•Review implementation notes

Step 2: Reopen and Start Timer

json

// Set back to in-progress
mcp__knowns__update_task({
  "taskId": "$ARGUMENTS",
  "status": "in-progress"
})

// Start timer (REQUIRED)
mcp__knowns__start_time({ "taskId": "$ARGUMENTS" })

Step 3: Document Reopen Reason

json

mcp__knowns__update_task({
  "taskId": "$ARGUMENTS",
  "appendNotes": "🔄 Reopened: <reason>"
})

Common reasons:

•User requested changes
•Bug found in implementation
•New requirements added
•Missed acceptance criteria

Step 4: Add New Requirements

json

mcp__knowns__update_task({
  "taskId": "$ARGUMENTS",
  "addAc": ["New requirement 1", "Fix: issue description"]
})

Step 5: Update Plan (if needed)

json

mcp__knowns__update_task({
  "taskId": "$ARGUMENTS",
  "plan": "Previous plan + new steps:\n1. Original step (done)\n2. Original step (done)\n3. NEW: Address new requirement\n4. NEW: Fix reported issue"
})

Present updated plan and WAIT for approval.

Step 6: Implement and Complete

Follow normal task completion flow:

json

// Check new ACs as completed
mcp__knowns__update_task({
  "taskId": "$ARGUMENTS",
  "checkAc": [<new-index>],
  "appendNotes": "✓ Done: new requirement"
})

// Stop timer
mcp__knowns__stop_time({ "taskId": "$ARGUMENTS" })

// Mark done again
mcp__knowns__update_task({
  "taskId": "$ARGUMENTS",
  "status": "done"
})

When to Reopen vs Create New Task

Reopen Existing	Create New Task
Small fix/change	Major new feature
Related to original work	Unrelated work
Same context needed	Different context
Quick addition	Significant scope

Rule of thumb: If it takes < 30 mins and relates to original task, reopen. Otherwise, create new task with reference.

Creating Follow-up Task Instead

json

mcp__knowns__create_task({
  "title": "Follow-up: <description>",
  "description": "Related to @task-$ARGUMENTS"
})

Remember

•Always document reopen reason
•Start timer when reopening
•Add new AC for traceability
•Stop timer when done
•Consider if new task is more appropriate