verify
A Claude Code plugin (“opslane-verify”) that performs spec-driven verification: it interprets an acceptance-spec document, extracts acceptance criteria, runs one browser automation agent per criterion (Playwright) against a local dev server, then a judge step reviews screenshots/traces to produce pass/fail results with evidence (screenshots and session recordings).
Score Breakdown
⚙ Agent Friendliness
🔒 Security
The provided content does not describe TLS usage, secret handling, or dependency posture for the plugin itself; it only references Playwright and that Claude Code uses OAuth login. Because the workflow runs browser agents against a local dev server, it may have access to application data visible in the UI—users should ensure test environments and data exposure are appropriate.
⚡ Reliability
Best When
You have a local dev server and a stable, testable acceptance-spec format, and you want quick feedback with visual evidence for each acceptance criterion.
Avoid When
You need strongly controlled, deterministic testing at scale (e.g., large suites in shared CI environments), or you cannot safely run browser agents that interact with the app under test.
Use Cases
- • Validating web-app behavior against a written set of acceptance criteria before pushing changes
- • Generating automated evidence (screenshots/video/Playwright traces) for failed acceptance criteria
- • Reducing manual QA effort for regression checks driven by spec documents
Not For
- • Use as a general-purpose CI/CD system (README explicitly states “No CI. No infrastructure.”)
- • High-assurance security testing or compliance validation without human review (it’s an automated verification workflow, not a formal verifier)
- • Environments where you cannot run browser automation against a local dev server
Interface
Authentication
The README mentions Claude Code with OAuth login via `claude login`, but does not describe scopes or plugin-specific authorization beyond requiring that login.
Pricing
No pricing information found in the provided content.
Agent Metadata
Known Gotchas
- ⚠ Spec interpretation/planning may require the user to provide a testable spec and answer clarifying questions before execution.
- ⚠ Browser-agent runs can be non-deterministic if the app under test is flaky or uses timing-dependent UI behavior.
- ⚠ Parallel execution per acceptance criterion (if implemented) may stress shared dev-server state; without explicit isolation, runs may affect each other.
- ⚠ The workflow is evidence-heavy (screenshots/video/trace), which may be impacted by large sessions or environment limitations.
Alternatives
Full Evaluation Report
Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for verify.
AI-powered analysis · PDF + markdown · Delivered within 30 minutes
Package Brief
Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.
Delivered within 10 minutes
Score Monitoring
Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.
Continuous monitoring
Scores are editorial opinions as of 2026-03-30.