assess-ticket

Run a ticket relevance assessment that compares the issue's motivation against current code. The relevance dimension checks whether the problem was solved another way or the target system was removed, returning relevant, uncertain, or superseded verdicts.

Use drift analysis, which extracts file paths, module names, and API references from the ticket and verifies they still exist. It also examines commits since the ticket's last-updated date in affected areas to determine whether assumptions hold, partially drifted, or are severely outdated.

Yes, pass a mode argument of drift, relevance, progress, advisability, or complexity to run a single dimension. The default mode is all, which runs every dimension in order and skips advisability and complexity when progress is complete.

The assessment accepts issue URLs, shorthand references like #99 or issue 99, file paths, or plain text. When no source is given, it auto-resolves from the environment, and resolved URLs are persisted to the branch manifest for later sessions.

Non-baseline verdicts trigger a numbered list of follow-up actions such as updating the ticket to match the codebase, revising it for advisability concerns, closing it, or adding a triage comment. The agent waits for you to pick an option before executing anything.