Daily Standup, Strategy Cleanup, and Catch-Up Linking

This release makes daily execution easier to read and strategy cleanup easier to trust. Daily Standup is clearer, Initiative and Theme Merge make duplicate strategy safe to consolidate, and Catch-Up Linking turns unlinked work into reviewable attribution clusters.

The last release made InteliG better at deriving strategy from execution. This one closes the loop around what happens after the system sees the work.

Once InteliG can infer strategy and attribute work, teams need the daily operating views to stay readable. They also need a way to clean up duplicate strategy, move historical work into the right place, and keep the reporting math honest.

That is the shape of this release: a better standup board, safer consolidation tools, a review surface built for clearing attribution backlog, and a set of reliability fixes that make the numbers easier to trust.

Daily Standup

The standup board is easier to scan and less brittle.

• Sort toggle. Switch between “Most active first” and “A to Z by name” without a page reload.
• Inactive today footer. Contributors with no activity on the selected day move into a collapsible footer, keeping the main board focused while still showing the full team picture.
• Contributor lookup fixed. A query bug that could prevent the standup from loading has been fixed.

Strategy Merge

Strategy cleanup is now a first-class workflow. When the system or a reviewer sees duplicate initiatives or duplicate themes, InteliG can consolidate the records without losing the execution history underneath them.

• Initiative merge. Select two or more initiatives, preview exactly what will be consolidated, confirm with a name-check gate, and merge them into one surviving initiative.
• Work follows the surviving initiative. Commits, pull requests, action items, decisions, and discovery data are reassigned automatically during merge, so the strategy record stays connected to execution history.
• Theme merge. The same workflow now exists for themes. Review the child initiative counts, confirm the merge, and preserve dissolved themes in an audit trail.
• Possible duplicates badge. The Initiatives page now flags semantically similar initiatives, explains the rationale for each candidate pair, and sends reviewers directly into the merge flow.
• Duplicate discovery routes to merge. Discovery candidates that match existing strategy now route to Merge instead of a misleading Confirm path that could create parallel initiatives.

Catch-Up Linking

The linking workflow now has a real Catch-Up tab for historical attribution work.

Instead of reviewing a long list of unlinked commits and pull requests one row at a time, InteliG clusters unlinked work into decisions a reviewer can move through quickly.

• Clustered review. Unlinked commits and pull requests are grouped into SURE, GUESS, NEW, and UNCLEAR clusters.
• Backlog-scale attribution. Hundreds of unlinked items can now be cleared in minutes, which directly improves alignment and effort reporting.
• Surviving links stay visible. Merged items include a click-through to the surviving initiative, preserving traceability after cleanup.

Cognis Write Actions

Cognis can now propose and execute more Knowledge changes on your behalf, using the same human-review pipeline as the rest of the product.

• Action items. Cognis can complete, pin, unpin, delete, and update action items.
• Decisions. Cognis can update or delete decisions.
• Review before mutation. Each write still flows through propose, approve, and execute. Cognis can suggest the change, but nothing mutates until the reviewer approves it.

Findings Inbox

Findings now have a real inbox instead of a placeholder screen. Detection agents can write to one shared place, and teams can review what changed without hunting through individual agent surfaces.

• One inbox for every detection agent. Findings from across the agent fleet now land in a dedicated Findings page, grouped by day and filterable by severity.
• Deep links and triage actions. Every finding can be opened directly from its URL, acknowledged, dismissed, or cleared in bulk with “Mark all read.”
• Live unread badge. The sidebar now shows unread findings as a live badge, so new risks are visible without opening the page.
• Cognis can answer from findings. Cognis can now query the findings inbox directly, so you can ask about recent alerts in conversation and get an answer grounded in the same triage feed.

Reporting Accuracy and Reliability

• Work-hour inflation fixed. Strategy reporting now anchors work-hour estimates to actual working sessions from commit timestamps, preventing the inflated hour totals that could overstate engineering effort in a reporting window.
• Merge auditability improved. Dissolved themes and merged strategy records preserve enough history for reviewers to understand what happened after consolidation.
• Discovery confirmation repaired. Confirmed discovery candidates now preserve their target date correctly, keeping newly accepted initiatives visible in strategy views.

Why it matters

• Daily Standup is more useful as an operating view. Active contributors stay prominent, inactive contributors remain visible, and the board no longer fails on contributor lookup.
• Strategy cleanup no longer breaks attribution. Duplicate initiatives and themes are unavoidable when strategy is inferred from fast-moving execution. Merge workflows let teams consolidate the model while preserving commits, PRs, decisions, action items, and audit history.
• Attribution backlog is now practical to clear. Catch-Up Linking makes the review loop fast enough to handle historical commits and PRs, improving alignment metrics without manual spreadsheet work.
• Cognis is becoming an operator. The new write actions move Cognis beyond answering questions into proposing concrete Knowledge changes, while approval stays in the human workflow.
• Findings are now operational. Detection agents can surface risks continuously, and the Findings inbox gives teams a shared place to triage them.
• Work-hour fixes improve trust. Corrected work-hour math makes strategy dashboards easier to rely on.

Release confidence

• Scope. intelig-backend and intelig-frontend, main branch, window 2026-05-21 to 2026-05-22.
• Evidence. 30+ commits and 14 merged pull requests across both repositories, covering Daily Standup, initiative merge, theme merge, Catch-Up Linking, Cognis write actions, Findings Inbox, and reporting accuracy fixes.
• Validation. Backend tests passed with 1,796 tests, 0 failures, 0 errors, and 4 skipped. Frontend unit tests passed with 5 tests, and the production client build completed successfully.
• Confidence. High for Daily Standup, initiative merge, theme merge, Catch-Up Linking, Cognis write actions, and Findings Inbox. Medium-high for work-hour reporting accuracy because the regression is covered and the calculation path was corrected, but continued production observation is useful because real commit cadence varies by organization.