OwlAudit
OwlAudit provides Hash Chain audit ledgers and policy-driven HITL / HOTL / HOOL three-tier interventions. Human review tickets support pending states, escalations, timeout fallbacks, and decision provenance. Synchronous evaluation runs on connected NodalOS hooks; generic ingest logs records without blocking past executions.
OwlAudit
Compliance Audit & HITLAI Agent compliance audit and human-in-the-loop (HITL) governanceCommitted actions are verifiable and traceable; on connected NodalOS hooks, policy-triggered high-risk actions await human sign-off.
- Verifiable
- Recompute hashes and detect gaps on committed chains
- Real-Time Blocking
- Synchronously rejects actions matching HITL policies on NodalOS hooks
- Three Tiers
- HITL · HOTL · HOOL
Typical pains
- Agents call production APIs directly without human oversight or guardrails
- Post-incident forensics lack complete, tamper-evident audit trails
- Compliance policies fail to translate into concrete Agent behaviors
Core value
Human-in-the-loop (HITL) approval gate
High-risk operations and critical code changes trigger human review via policy interception; approval confirmations and override gates are managed in AULO and written back upstream.
Human review statistics for each project
OwlAudit tracks each project's human review work items — approval rate, average resolution time, human coverage and escalation rate — and lists every item with its category and outcome. They appear on the audit tab of the project page in AULO.
Audit policies and compliance summary
OwlAudit's compliance summary — audit entries, decisions, human review items and approval rate, audit chain integrity and the most-triggered policies — sits in the audit settings of the AULO admin console, next to each audit policy's trigger event and action. Each policy can be switched on or off.
Capabilities
Audit ledger
Human-in-the-loop
Policy & freeze
Reporting & integration
Does / does not
- Verifiable Hash Chain audit ledger: recompute, verify, and detect chain gaps
- Three-tier HITL / HOTL / HOOL intervention: policy-driven human-in-the-loop gates
- Policy dual-version provenance: policy changes and audit records increment versions in the same transaction
- Human review ticket lifecycle: pending, escalation, timeout fallback, and decision provenance
- Compliance freeze and real-time blocking: synchronous deny and freeze on connected NodalOS hooks
- Objective audit reporting and veto metrics: compliance summaries and objective stats for AULO
- Does not replace professional legal and compliance advisory
- Does not perform subjective black-box scoring (records facts and human veto rates, not unexplained scores)
- Does not replace existing enterprise GRC / SIEM suites (focuses on Agent runtime control and in-product compliance reports)
- Does not perform physical off-chain real-world investigation (verifies integrity of records already written to the ledger)
By role
Works with NodalOS hook policy evaluation to reject and log high-risk actions
Provides a complete loop across reviews, escalations, and audit trails rather than mere pre-call blocking; tracks review veto frequency for governance
On connected hooks, policy-triggered HITL actions are blocked synchronously pending sign-off