In-flight work from 2026-06-05 that was sitting uncommitted on
feat/agent-pipeline-bridge for 10 days. Moved here to a clean branch
off main so the bridge branch can stay at its committed state (the
DEFERRED Phase 2a snapshot, see stwl-labs/ubiquitous-invention#3).
Touched areas:
- apps/web/app/(app)/[workspaceSlug]/settings/runs/page.tsx — runs page UI iteration
- apps/web/server/routers/{backlog,runs}.ts — router updates
- apps/web/app/(app)/[workspaceSlug]/plans/[planSlug]/[epicSlug]/[taskSlug]/page.tsx — new deep-link route
- apps/web/components/backlog/workflow-prompt-section.tsx — new component
- config/CursorSync.md — coordination doc tweak
- docs/operations/README.md + 2026-06-05-db-drift-reconcile-oauth-tables.sql — operations ops note + reconcile script
- plans/Plan-agent-coordination/Epic-task-as-runnable-unit/Task-{workflow-prompt-task-detail-ui,deep-link-runs-to-task-detail}.md — 2 task specs
This is a working-tree snapshot, not a finished PR. Rebase, split, or
amend as needed when picking it back up.
Co-authored-by: Cursor <cursoragent@cursor.com>
Adds level-B agent attribution: every MCP tool call can identify itself
with a `client` (e.g. cursor-ide, codex, echodo-orchestrator) and `model`
(e.g. claude-4.6-sonnet) string. Both optional, both stored in
agent_runs.metadata as JSONB so the schema doesn't move.
- claim_task: new optional `client` + `model` args. Written into the
inserted agent_runs row's metadata, and mirrored into the audit log
entry. On idempotent re-claim, incoming values are merged into
existing metadata (existing keys override only when explicit), so a
mid-session model switch updates attribution without losing earlier
context.
- complete_task: same optional args, with merge-on-close semantics —
late-bound values override the earlier claim's values so the run row
reflects whichever model actually closed the session. Audit row also
carries the merged identity.
- /settings/runs UI: stack the client/model under the actor name in the
table so attribution is visible at a glance without adding a column.
Smoke-tested: claim with {cursor-ide, claude-4.6-sonnet} then complete
with only model={claude-4.6-sonnet-medium-thinking} yields final
metadata {client: cursor-ide, model: claude-4.6-sonnet-medium-thinking}
on both the run row and the task.completed audit entry.
Co-authored-by: Cursor <cursoragent@cursor.com>
Read-side only — write paths land with claim_task / complete_task
in the next epic. Keyset pagination on started_at, three procedures
(listRecent, listForTask, summary), and a /settings/runs view that
mirrors the audit page's visual language.
Co-authored-by: Cursor <cursoragent@cursor.com>
