Agent Control Plane — seat-lensed (wf3)

wf3 — fully seat-lensed: the lead (purpose/KPIs/attention/actions) re-orders by ?seat (Finance→spend-at-risk, Security→control coverage, Exec→governance index, Eng→instrumentation/latency, Compliance→frameworks); roster gains governed? + compound-risk badge.

wf3 — leads with Control Coverage % + compound-risk (Critical & ungoverned) and a fleet trajectory (weeks-to-target).

wf3 — adds forecast-vs-budget + runway, provider concentration, cache savings, and governance-tagged (at-risk) spend.

wf3 — adds trajectory (weeks-to-target), a Critical/Elevated/Standard governance breakdown, and shadow-AI financial exposure.

wf3 — adds instrumentation coverage % and a single "fix next" worst-agent line (p95 + cost + governed) instead of separate counts.

wf3 — adds conformance coverage (frameworks-with-evidence, flags zero-finding = false confidence) and high-findings-by-agent accountability.

wf3 — seat-lensed (Security: gaps, risk-first sort; Engineering: utilization, cost sort); compound-risk badge, risk-factor cell, shadow-age hint.

wf3 — adds "why this score" (factors × points), delta vs worst_score, data-risk sub-score, MCP footprint, compound-risk banner; seat-aware lead emphasis.

wf3 — governance-tags the spend (governed/risk per agent + "$X at-risk/ungoverned" callout) and adds cost-per-session + $/1k-tokens-by-model; light exec lens.

wf3 — adds liability $ (spend with no chargeable owner), chargeback-readiness %, and an explicit risk-adjusted ranking note.

wf3 — adds $/call (slow AND expensive), latency-by-model, and variance (p99−p50)/avg; Task Outcomes kept as the named projected gap.

wf3 — seat-lensed (Security: ungoverned-agent findings jump the queue; Compliance: group by framework); score-impact (+pts if fixed) per mapped finding.

wf3 — "what lowers the score" becomes a Hardening-ROI fix-this-first roadmap; adds effective-posture worst-case (price the shadow gap) and a posture trend.

wf3 — seat-lensed (Compliance: privacy stake + owner; Security: credential/secret paths × agents); per-agent high-severity exposure drill + read/write asymmetry.

wf3 — per-agent actionable dry-run (names, not counts), approval-policy coverage %, tool-call diversity, and "cost of flipping to Enforce".

wf3 — adds the inference-model line (a finding reveals a gap, not a failed control), tags ungoverned-agent findings, audit-readiness depth/breadth, and evidence-gap (shadow-only) flags.

wf3 — adds compound badge (weak cred on a dangerous agent), last-used ("safe to rotate now"), credential-age (>90d), and org-binding coverage %.

wf3 — per-connector payload estimates and a shadow-agent export-gap warning (exports inherit the unmonitored blind spots).

wf3 — adds the governance-vs-spend story line, at-risk-spend concentration (top-3 share), a projected findings-backlog trend, and a benchmark caption on the index.