todo-tracker

name	todo-tracker
description	Automatic extraction of tasks, decisions, open questions, and blockers at the end of each session. Triggers on "TODO", "to do", or alongside a handover. Scans the entire conversation using a 6-category checklist to ensure nothing is missed. Outputs a human-readable summary + structured JSON.

Skill: TODO Tracker

Version: 1.0
Part of: never-start-over
Companion skills: handover (project state), retex (error post-mortem)

---

1. Mission

When the user asks for a TODO:

1. Scan the ENTIRE conversation (not just the last 10 messages)
2. Extract EVERYTHING that deserves tracking
3. Generate a human-readable summary + structured JSON
4. Assign ownership, priority, and attack order

---

2. Extraction Checklist (anti-forgetting)

The AI MUST review these 6 categories:

Note: Trigger phrases below are in English. Adapt to the conversation's language — the patterns matter, not the exact words.

A. EXPLICIT TASKS

Trigger phrases:

• "We need to...", "I'll...", "You should...", "Next step..."
• "Create...", "Modify...", "Test...", "Verify..."
• "Don't forget to..."

B. DECISIONS MADE

Trigger phrases:

• "OK let's go with...", "Validated", "We'll keep..."
• "Option X wins", "Final answer..."
• Any explicit commitment to an approach

→ Becomes a task IF it requires a follow-up action.

C. OPEN QUESTIONS

Trigger phrases:

• "To be verified...", "To be confirmed..."
• "I'm not sure yet...", "Need to check..."
• "?" without a clear answer in the conversation

D. MENTIONED BUT UNRESOLVED TOPICS

Indicators:

• Topic mentioned then conversation drifted
• "We'll come back to this", "Worth exploring"
• Topic without an explicit conclusion

E. BLOCKERS

Trigger phrases:

• "Doesn't work", "Bug", "Error"
• "Blocked by...", "Waiting for..."
• "Can't move forward without..."

F. INSIGHTS / DISCOVERIES

Trigger phrases:

• "I just found out...", "Actually..."
• "Good news", "Bad news"
• New information that changes the game

---

3. Ownership

Tag	Meaning	Example
owner: user	User handles it	"Test the script on mobile"
owner: ai	AI handles next session	"Code the dashboard"
owner: external	Depends on a third party	"Wait for IT response"
owner: both	Collaboration	"Validate specs together"

---

4. Priorities

Level	Label	Criteria
P0	Critical	Blocking, urgent, deadline approaching
P1	Important	Necessary but not urgent
P2	Normal	Do when possible
P3	Nice-to-have	Optional, improvement

Priority signals:

• "ASAP", "urgent", "critical" → P0
• "This week", "Soon" → P1
• "When you can", "At some point" → P2
• "Would be nice", "Bonus" → P3

---

5. JSON Schema

{
  "project": "PROJECT_NAME",
  "generated_at": "2026-04-17T18:30:00",
  "session_uri": "conversation-link-or-id",
  "summary": "2-3 sentence session summary",
  "tasks": [
    {
      "id": "PROJECT-001",
      "text": "Clear task description",
      "priority": "P0",
      "status": "open",
      "owner": "user",
      "category": "action",
      "tags": ["tag1", "tag2"],
      "context": "Why this task exists (1 line)",
      "deadline": null,
      "created_at": "2026-04-17",
      "done_at": null
    }
  ],
  "decisions": [
    {
      "text": "Going with local Python server",
      "reason": "Enables dynamic folder scanning"
    }
  ],
  "open_questions": [
    "How to handle merge conflicts?"
  ],
  "attack_order": ["PROJECT-001", "PROJECT-003", "PROJECT-002"],
  "next_session_focus": "Code the server + dashboard"
}

---

6. Task Categories

Category	Description
action	Concrete task to execute
decision	Decision requiring follow-up
question	Question to resolve
blocker	Blocking problem
insight	Discovery to retain
idea	Idea to explore later

---

7. Naming Convention

TODO_{{PROJECT}}_{{YYYYMMDD}}_{{HHhMM}}.json

---

8. Storage

Save the JSON wherever your project stores session data. Suggested locations:

• .session/ directory in project root
• A dedicated todos/ folder in your knowledge base
• Same directory as your HANDOVER files for co-location

---

9. Cross-Session Review

At the start of a new session, if a previous TODO exists:

1. Load the previous TODO JSON
2. Review each task with the user: done, still open, or dropped
3. Mark completed tasks ("status": "done", "done_at": "date")
4. Carry forward open tasks into the new TODO
5. This prevents task drift — things that were "P1 important" don't silently disappear

---

10. Output Format

When the user triggers "TODO":

1. Human summary (3-5 lines)

What was done, key decisions, next steps.

2. Extracted tasks (readable table)

#	Task	Priority	Owner
1	...	P0	user

3. Open questions

List if applicable.

4. Attack order

Recommended sequence for tackling the tasks.

5. Full JSON

Code block with JSON ready to save.

---

11. Anti-Patterns (do NOT extract)

Pattern	Why
"We should have..."	Past, not actionable
"It would be nice to..." (vague)	Too fuzzy — ask for precision
"I did X"	Already done
Off-topic conversation	Not project-related
Jokes, digressions	Not relevant

---

12. Verification Checklist

Before generating the JSON:

• Re-read the ENTIRE conversation?
• Captured explicit tasks?
• Captured decisions made?
• Any unanswered questions?
• Any topics mentioned then dropped?
• Any blockers mentioned?
• Every task assigned an owner?
• Priorities assigned?
• Attack order is logical?

---

13. Quality Criteria

A TODO extraction is good when it passes these tests:

Completeness test: Read through the conversation one more time after generating the TODO. Did you miss any action item, question, or blocker? If you find even one, the extraction failed.

Actionability test: Could someone execute each task from its description alone? "Fix the bug" fails. "Fix the race condition in WebSocket reconnect logic (server.py line 47)" passes.

Ownership test: Every task has exactly one owner. "Both" is acceptable only for tasks requiring real-time collaboration. If unsure, assign to the user — they can reassign.

Priority test: Do the priorities reflect the conversation's actual urgency signals, not just the AI's guess? Re-read the user's exact words before assigning P0.

Attack order test: Does the sequence respect dependencies? A task that depends on another must come after it. Does it front-load blockers?

---

Why This Skill Exists

At the end of a productive AI session, there are always tasks scattered across the conversation — some explicit ("we need to test this"), some implicit (a decision that requires follow-up), some invisible (a question asked but never answered). Without systematic extraction, these tasks fall through the cracks.

The 6-category checklist ensures nothing is missed, and the attack order transforms a flat list into an actionable plan.