Skip to content
Back to skills

Judge Task Lifecycle

ASecurity

Judge how a memory changes active-task state and whether it refers to an existing task. Use for mem-claw task-lifecycle model calls.

  • 984 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 22, 2026
ai-agents

Works with

  • cli

Security analysis

A100/100

Scanned September 22, 2026

npx -y skills add sno-ai/sno-station --skill judge-task-lifecycle --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Judge Task Lifecycle?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Judge Task Lifecycle
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/sno-ai-judge-task-lifecycle/badge)](https://www.skillsdirectory.com/skills/sno-ai-judge-task-lifecycle)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: judge-task-lifecycle
description: Judge how a memory changes active-task state and whether it refers to an existing task. Use for mem-claw task-lifecycle model calls.
---

# Judge Task Lifecycle

Judge task meaning. Do not decide from exact wording alone.

## Active Task State

Apply the `To-do Boundary` section supplied from `extract-atomic-memory` before these state rules.

Decide in this order and stop at the first step that applies.

1. **Does the candidate report OR INSTRUCT that an existing active task is done, finished,
   handled, cancelled, or dropped?** If yes, the answer is `complete` or `remove` on THAT task.
   Never `open_or_refine`, and never `none`.
   A report ("I finished X", "X is done") and an instruction ("mark X complete", "tick X off",
   "close X", "remove X from my list") both land here. An instruction to change a task's state
   IS the state change; do not treat it as a request you decline to act on.
   A candidate that speaks about an active task's ending is never a new task, however it is worded,
   and restating the task's name inside such a sentence is naming the task to close, not opening one.
2. Otherwise, if the candidate describes work still to be done, use `open_or_refine`.
3. Otherwise use `none`.

- Use `open_or_refine` only for a current task that remains active.
- Use `complete` only when the candidate says an active task was achieved.
- Use `remove` only when the candidate says an active task was cancelled or should no longer be tracked.
- An explicit instruction to mark a task complete, done, or finished is `complete`, even when it is
  phrased as an edit to a list — "mark X complete on my todo list" and "tick X off" are `complete`,
  not `remove`. Taking a finished task off a list is what completing it looks like; the words that
  decide are whether the work was done. Reserve `remove` for a task abandoned, cancelled, called off,
  no longer needed, or added by mistake — cases where the work will NOT happen.
- Use `none` when the candidate does not change active-task state.
- A candidate that reports a task as finished, cancelled, or already handled never opens one.
- If that task is active, complete or remove it. If it is not active, make no task change.
- Match tasks by meaning, not exact wording. Refine an existing task instead of opening a duplicate.

## Existing Task Relation

- Use `same_instance` when the assertion refers to one listed active task, including different wording with the same meaning.
- Use `distinct_instance` when the assertion is a different active task.
- Use `none` when the assertion does not describe an active-task relation.
- Use `uncertain` when the evidence does not settle the relation.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.