Week of June 01, 2026
Delivery started keeping its own paper trail this week: Changes now push status, summary, and description updates straight to linked GitHub and Jira items automatically. PR cards got AI summaries and capability suggestions, and the Capability Landscape gained a scrubbable history timeline. Highlight: Changes sync automatically to GitHub and Jira, closed-but-unmerged PRs included.
In this edition
Changes sync itself to GitHub and Jira
Changes stopped being a one-way record of what shipped. They now sync automatically to their linked GitHub and Jira items, pushing status, summary, and description updates outward and transitioning the external workflow state to match. Status mapping got smarter too: a closed-but-unmerged GitHub PR now shows as Cancelled instead of quietly looking like Done, and failed or rolled-back changes map to the right cancel state in Jira. If background sync fails to start, a clear toast tells you immediately instead of leaving you to wonder.
Read moreRead less
PR change cards got smarter at the same time. Instead of raw PR text, cards and detail views now show an AI-generated summary when one's available, alongside capability suggestions you can accept by linking an existing capability or creating the one it proposes. Manual controls let you re-run a suggestion or reject it outright, so automation stays a suggestion, not a decision made for you.
- Changes auto-sync status, summary, and description to linked GitHub and Jira items
- Smarter status mapping: closed-but-unmerged PRs show as Cancelled, failed/rolled-back changes map to the right Jira cancel state
- Clear toast warning when background sync fails to start
- AI-generated PR summaries in place of raw PR text on change cards and detail views
- Capability suggestions on PR changes, acceptable by linking or creating a capability
- Manual re-run and reject controls for AI suggestions
The Capability Landscape gets a timeline
The Capability Landscape picked up a compact, responsive timeline this week: pick a date from a calendar, jump to now, or play forward to step through a capability's lifecycle. Historical snapshots show capabilities, including ones you've since soft-deleted, exactly as they existed on the day you're looking at. Opening today's view still gives you an editable workspace, opening a historical snapshot from today opens read-only, so you can't accidentally edit history.
Read moreRead less
Import got a matching upgrade. A unified CSV/JSON import flow, with a minimal JSON format of just name, description, and parent, previews the proposed graph and list changes before anything commits, lets you selectively apply updates, restore capabilities you'd previously deleted, and undo the last import if something looks off. Imports validate integrity first, blocking unresolved parents and parent cycles, and roll back automatically if a partial import fails.
- Compact, calendar-driven timeline: pick a date, jump to now, or play forward through capability lifecycles
- Historical snapshots show capabilities (including soft-deleted ones) as they existed at a selected time
- Today's view opens editable, historical snapshots from today open read-only
- Unified CSV/JSON import with a minimal JSON format (name, description, parent)
- Import preview, selective apply, restore of previously deleted capabilities, and undo of the last import
- Integrity validation blocks unresolved parents and parent cycles, with automatic rollback on partial failure
Feedback and discovery tighten up
A handful of consistency fixes landed in Feedback and discovery. Theme detail cards now show chunk-level status for linked evidence, so a theme's card and the underlying feedback agree about where things stand. Actions like linking, merging, creating a capability, re-analyzing, or changing status now reliably trigger updates to the feedback graph instead of leaving it stale. Hypothesis creation also moved into the Validation Tree itself: the standalone "New Hypothesis" button is gone, and hypotheses now get created from the tree's context menu, tied to the Bet they belong to from the start.
Read moreRead less
- Theme detail cards show chunk-level status so theme, evidence, and feedback stay consistent
- Feedback graph updates reliably after linking, merging, capability creation, re-analysis, or status changes
- Hypotheses are created from the Validation Tree context menu, tied to their Bet, replacing the standalone "New Hypothesis" button
Agents and workspace get quieter
A few quieter changes rounded out the week. Vega and Polaris stopped restarting cold: conversations now open with compact strategy and discovery context, and both agents hold onto lane-specific workspace context as you move between turns. Workspace Settings gained a new Testlab section that centralizes experimental toggles, off by default, for Discovery Learnings, External Connections, and the Feedback Graph, so teams opt into only what they want. A handful of visual fixes tidied up avatars, chat bubbles, and the usage panel along the way.
Read moreRead less
- Vega and Polaris open conversations with compact strategy and discovery context up front, and retain lane-specific workspace context across turns
- New Testlab section in Workspace Settings centralizes experimental/beta toggles for Discovery Learnings, External Connections, and the Feedback Graph, off by default
- Unsaved-changes warning prevents losing Testlab settings when navigating away
- Consistent avatars and compact chat bubble rendering across Vega, Polaris, crew roster, and launcher surfaces
- Cleaner chat usage panel layout with a redundant divider removed
What this means for you this week
- Changes and their linked GitHub/Jira items stay in sync automatically, with accurate status instead of manual updates.
- PR reviews come with an AI summary and capability suggestions you can accept, reject, or re-run.
- Step through a capability's full history on a real timeline, and import or restore capabilities with a safe preview.
- Vega and Polaris pick up conversations without losing context, and Testlab lets you turn experimental features on or off centrally.
If you want a walkthrough of the GitHub/Jira sync, the Capability Landscape timeline, or the new Testlab settings, reply here and I'll set up a quick demo.