▦ WORKSPACE
IDE Runtime
SGT
Commander StatusStanding By

No active mission.

Current MissionRepository Review

Mission ready. Awaiting orders.

Current Workspacesergeant

Branch: —

Commander VerdictWAITING

No runtime evidence yet.

COMMANDER VIEW

Mission → Officers → Armoury → Evidence → Verdict

EvidenceVerificationReportVerdict

SERGEANT

Professional. Disciplined. Relentless.
Delivering verified code reviews with deterministic proof and evidence-grounded open models.

SGT
COMMAND
Read-only defaultHuman approvalNo auto-merge

Recent Missions

Confidence

IDE Awareness

Active file—
Changed files0
PlatformIDE
Semantic routeAutomatic

Officer Deployment

Mission Planner

Mission Type → Briefing → Permissions → Loadout → Assigned Officers → Weapons Selected → Launch Mission.

Mission Type

Mission Loadout

Semantic Review Router

Automatic discovery is loopback-only. Remote code leaves the machine only when you explicitly set a remote Base URL. API credentials stay in SERGEANT_LLM_API_KEY and are never stored in this webview.

Pass to Writer

Writer ModeDisabled by default
Patch OutputDraft patch only
ApprovalHuman approval required
Auto MergeNever auto-merge

Mission Permissions

Mission Timeline

✓ Mission staged
2 Officer deployment
3 Commander verdict

Mission In Progress

Commander is deploying officers, selecting weapons, and gathering deterministic and semantic evidence.

Mission Timeline

0%

Current Operation

Waiting for runtime…

Armoury Selected

Static AnalysisACTIVE
Semantic ReviewROUTED
Regression TestsACTIVE
Battle CompareQUEUED


    
Evidence
Static, runtime, semantic, UI, documentation, battle and external evidence.


    
Evidence Locker
Latest Reports · Mission Reports · Battle Reports · Verification / Release Reports · Search · Export · Copy · Save.
Latest Report
Battle Reports
Branch comparison, UI proof and before/after evidence.
Report Actions
OrderDateResultMissionDuration


    
Officer System / Armoury
Commander → Officers → Armoury → Evidence → Verdict
Weapon Manifest
Officer Deployment


    
Settings
Grouped for extension readiness.


    
Sergeant V2 Review Doctrine
Evidence first. Verdict second. Nothing is assumed. Sergeant gathers comparable evidence, then the Commander reasons from that evidence.
Static Evidence▼Runtime Evidence▼Semantic Evidence▼UI Evidence▼Docs Verification▼Cross Verification▼Confidence▼Commander Verdict
Benchmark Loop
  1. Design evidence weapon
  2. Implement and validate
  3. Review implementation
  4. Compare independent models and external reviewers
  5. Investigate every disagreement
  6. Improve evidence and reasoning


    
Post‑V2 Roadmap
Future enhancements remain separate so V2 stays focused on Commander → Mission → Officers → Weapon Manifest → Evidence → Verdict → Audit Trail.


    
◇ What is Sergeant?
Documentation for the Commander / Mission / Officer architecture.

  


  
Quick Actions