v0.29.0: Implementation trails, Work reports, and A real setup wizard for chat intake
v0.29.0 is a big one — two brand-new planning and reporting workflows, a proper setup wizard for chat intake, and Opus 5 / Sonnet 5 wired in as defaults.
Implementation trails
Big features rarely fit in one task, and figuring out what to do first across a dozen related tasks is a chore you end up redoing every week. /aitask-trail does that analysis once and stores the answer: a wave-by-wave implementation plan, backed by evidence from the actual task files, plans, and gate state. It lives as a versioned artifact, so it survives the session — and when the underlying tasks change, the trail tells you it’s stale and why. The board picks it up with a new By-Trail view (z): waves become columns, each card shows its classification and confidence, and enter opens the full narrative.
Work reports
Someone asks what you shipped this week and you go spelunking through git log. Not anymore. Hit w on the board, pick the columns you care about, and /aitask-work-report drafts a manager-facing summary: what landed, what’s in flight, what’s queued. Turn on the projection and it’ll estimate a delivery date from your actual recent velocity — with a confidence floor, so it stays quiet when it doesn’t have enough data to be honest.
A real setup wizard for chat intake
Configuring the Discord bug-report intake used to mean hand-editing YAML and guessing at channel IDs. The ait chatlink TUI now has a config-check panel showing exactly what’s working and what isn’t, plus a w wizard that walks you through the whole thing. It validates your bot token against Discord live — login, privileged intents, channel visibility, permissions — then fetches your channel members and roles so you pick them from a list instead of pasting snowflake IDs. Get interrupted? It saves a token-free draft and offers to resume.
Sharper shadow reviews
The shadow companion’s implementation review now comes in four tiers — Quick, Basic, Standard, Deep — so you can ask for a hunk-level sanity check or a full adversarial interrogation. Every finding carries a disposition (blocking, follow-up, or informational) and a verdict, and the rules were reworked so a real defect can never be quietly dropped to stay under a cap. If something is omitted, the review has to say so.
Opus 5 and Sonnet 5
Both are registered for Claude Code. Opus 5 is now the default for the heavy operations — pick, explore, learn, trail, and brainstorm — and Sonnet 5 takes over the lighter ones.
Full changelog: v0.29.0 on GitHub