AI Review Gate

AI review evidence for scope, security, dependencies, CI, and payout readiness.

The AI gate reports evidence and manual review areas. It never claims absolute safety or final authority; it helps sponsors decide whether a bounty can move toward acceptance.

Loading
Audited claims

Competing claim reports scored by the specialized audit agent.

Loading
Queued gates

PR gates waiting for claim evidence, AI review, or acceptance.

Loading
Trust score

Sponsor reliability remains visible beside review evidence.

Review reports

Real AI review summaries attached to your current Convex bounties.

Loading reports

LockBounty is reading review evidence from Convex.

Independent review checks

Viktor and Kael provide evidence and recommendations. They do not replace the sponsor or maintainer's final acceptance decision.

Viktor
Primary PR auditor

Reads the pull request, gathers CI, dependency, and security evidence, then recommends a bounded LockBounty report.

Waiting
Kael
Adversarial counter-review

Challenges Viktor's ready verdict, highlights missing proof, failed checks, or high-risk security findings.

Waiting
Bridge
Hermes handoff timer

Keeps the Viktor to Kael handoff alive, retries pending reviews, and emits health pulses.

Waiting

Signed agent activity

Recent Viktor, Kael, and bridge signals accepted by LockBounty. Manual smoke checks are not payout authority.

Loading activity

LockBounty is reading the latest Hermes pulses from Convex.

Report schema

The evidence structure used to keep AI recommendations bounded, reviewable, and non-binding.

Scope match

Compare acceptance criteria to changed files and PR summary.

Evidence
Security review

Flag secrets, auth, payment, dependencies, and unsafe execution paths.

Risk
Code cleanliness

Rate maintainability, clarity, test evidence, performance, and CI status.

Score