completion verifier

The problem

Task systems rot when items sit in a "probably done" limbo. You waste mental cycles and planning time routing around tasks that were started or submitted but never actually verified as finished in the real world.

What it does

• Interrogates "I did it" claims to distinguish between actual completion and assumed completion.
• Classifies evidence into three tiers based on hard data, soft memory, or unverified status.
• Generates structured Verification Briefs for unconfirmed tasks to settle status in under two minutes.
• Runs automated "confirmation sweeps" at the end of work sessions to catch unlogged status changes.
• Forces unverified tasks to remain on the active list, preventing them from slipping through the cracks.

Why this beats prompting it yourself

A standard LLM will take your word for it when you say a task is finished. This skill is programmed to be professionally skeptical, requiring specific evidence types like confirmation numbers or timestamps before clearing a log. It converts vague user intent into a repeatable verification protocol that prevents project debt.

Use cases

• Verifying form or application submissions in external portals.
• Confirming email delivery or client receipt of deliverables.
• Auditing a daily task list before logging off to ensure zero "limbo" items.
• Checking if automated processes or deployments actually cleared.