Aegis: Closed-Loop Intelligence Engine
Ground behavior, improve it, and defend every ship decision with evidence.
Mode
Eval-first
Release
Gate-aware
Reports
Shareable
Access
Accounts enabled
Public siteWorkspace settings
Shell policy
The workspace chrome does not inject sample benchmark rows, synthetic scores, or decorative regression traces. Live evidence belongs in the closed loop, research runs, review queue, and release train after a real workspace is populated.
Closed Loop
Import traces, run the strict loop, and open the dossier.
Research Runs
Measure benchmark deltas and investigate candidate behavior.
Review Queue
Attach ownership, severity, and operator judgment.
Release Train
Persist gate state beside the same artifact lineage.
Launch-grade proof should be grounded in persisted artifacts, not shell placeholders.
surface
purpose
required
owner
dataset
fixed benchmark contract
yes
research
comparison
baseline vs candidate delta
yes
operator
review
annotated release judgment
yes
human
promotion
gate outcome + lineage
yes
release
Human gate

Triage evals and strict dossiers, capture operator judgment, and decide what actually ships.

Use one queue for both sides of the product loop: experiment review and closed-loop run review, with the same decision language carried into release train.

Open release trainOpen closed loopCompare runs
Visible items
0
Needs attention
0
Approved
0
Filters
Narrow the operator inbox by source and status before you start assigning judgment.
Queue
Items waiting for review
Loading review queue...
Review details
Select an eval run or closed-loop dossier from the queue to review it.