VerifAIer Sentinel · AI Governance Infrastructure · Horizon 2
Deep Governance Audit
Submit any AI session transcript and run the full Horizon 2 oversight pipeline:
reasoning cascade analysis, temporal trust scoring, authority chain governance, and per-turn stage classification.
Deterministic. Local-first. Independent of model provider. Replay-safe.
Sentinel API offline.
The "Analyze" button requires vai serve running on localhost:8000.
To start it: open a terminal in your VerifAIer project and run python -m vai.sentinel_server or vai serve.
In the meantime, use Load Example Session to see a pre-analyzed showcase result.
Session Transcript
Paste session JSON or load the executive showcase example.
Running full governance pipeline: cascade, temporal, authority analysis…
authoritative_for_wiw = false —
All findings are advisory governance signals. VerifAIer does not determine who wrote code,
who made a decision, or who is legally responsible for any governance failure.
Role classification is heuristic pattern-matching, not semantic inference.
src/auth.py:10 severity: HIGH session: ses-8f2a1b3c
02 Escalation
ESC-2026-05-16-0031
auto-created status: PENDING REVIEW sla: 24h
03 Reviewer
Assigned: security-lead
vai escalation assign ESC-2026-05-16-0031 --to security-lead
04 Approval
APPROVED TO REMEDIATE
vai escalation approve ESC-2026-05-16-0031 --note "rotate key, no exposure"
05 Remediation
RESOLVED
Key rotated New receipt sealed Audit trail updated
# Full escalation lifecycle — copy-paste ready
vai escalation list --status pending ESC-2026-05-16-0031 HARDCODED_SECRET HIGH ses-8f2a1b3c PENDING REVIEW
vai escalation assign ESC-2026-05-16-0031 --to security-lead
vai escalation review ESC-2026-05-16-0031
vai escalation approve ESC-2026-05-16-0031 --note "key rotated, no external exposure confirmed"
vai escalation close ESC-2026-05-16-0031 --resolution remediated