Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation
The English summary is being prepared. Showing the original description.
Platform design review critique dimensions and severity levels. Load when reviewing CI/CD pipelines, infrastructure, deployment strategies, observability, or security designs.
The English summary is being prepared. Showing the original description.
See how to installReview the instructions the agent will receive before you install.
Questions: All stages defined (commit, acceptance, capacity, production)? Quality gates explicit with pass/fail criteria? Parallelization used? Failure recovery/retry documented? Commit stage < 10 min, acceptance < 30 min?
Blocker: Missing critical stage (no acceptance tests) | no quality gates | no security scanning. Critical: Pipeline > 30 min without parallelization | no failure notification | missing artifact versioning. High: No caching strategy | incomplete environment parity | missing matrix testing. Medium: Inconsistent naming | missing documentation for manual steps.
Questions: Infrastructure fully codified? Modules reusable and parameterized? State management secure (encrypted, locked)? Security best practices (least privilege, encryption)? Idempotent and reproducible?
Blocker: Secrets in version control | no state management | production credentials in code. Critical: No encryption at rest | overly permissive IAM | missing network security. High: Hardcoded values | missing resource tagging | no cost estimation. Medium: Inconsistent module structure | missing input variable validation.
Questions: Strategy appropriate for risk profile? Rollback documented? Health checks and readiness probes defined? Gradual traffic shifting with automatic rollback? Database migrations backward compatible?
Blocker: No rollback strategy | no health checks | breaking changes without safeguards. Critical: Single-shot deployment for critical services | no canary/blue-green for high-traffic | missing pod disruption budgets. High: Rollback not tested | no gradual traffic shifting | no pre-deployment validation. Medium: Incomplete manual step documentation | no feature flags for risky features.
Questions: SLOs defined with specific targets? All four golden signals monitored? Distributed tracing configured? Alerts SLO burn-rate based? Dashboards for investigation?
Blocker: No SLOs defined | no error rate monitoring | no alerting strategy. Critical: No latency monitoring (p50/p90/p99) | symptom-based alerts | no log-metric-trace correlation. High: Incomplete metric coverage | alert thresholds misaligned with SLOs | no runbook links. Medium: Unclear dashboard organization | missing error budget tracking.
Questions: SAST in commit stage? DAST before production? SCA configured? Secrets management using external vault? SBOM generated and signed?
Blocker: No security scanning | secrets in env vars or code | no container image scanning. Critical: Missing SAST in CI | no dependency vulnerability scanning | missing K8s network policies. High: No DAST before production | no SBOM generation | no image signing. Medium: Security scan results not blocking deployment | no license compliance.
Questions: Design enables multiple deployments/day? Lead time < 1 hour achievable? Change failure rate tracking in place? Time to restore measurable with SLOs?
Critical: Manual steps preventing daily deployments | no automated testing for fast feedback | no deployment failure tracking. High: Pipeline > 1 hour for full deployment | no post-deployment validation | missing deployment frequency metrics.
Questions: Design addresses largest bottleneck first? Simpler alternatives documented and rejected with evidence? Constraint prioritization correct? Complexity justified?
Critical: Design addresses secondary concern while larger exists | no measurement data | simple alternatives not documented. High: Constraint prioritization not explicit | over-engineered for stated requirements.
No reviews yet. Tell others what you thought.
No login needed. Share how it worked for you, ask a question, or note where it did not run.
No comments yet. Be the first to write one.
Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation
The English summary is being prepared. Showing the original description.
Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation
The English summary is being prepared. Showing the original description.
Review dimensions for acceptance test quality - happy path bias, GWT compliance, business language purity, coverage completeness, walking skeleton user-centricity, priority validation, observable behavior assertions, traceability coverage, and walking skeleton boundary proof
The English summary is being prepared. Showing the original description.
Detailed 5-phase workflow for creating agents - from requirements analysis through validation and iterative refinement
The English summary is being prepared. Showing the original description.
5-layer testing approach for agent validation including adversarial testing, security validation, and prompt injection resistance
The English summary is being prepared. Showing the original description.
Architectural style selection decision matrices, trade-off analysis, structural enforcement rules, and combination patterns. Load when choosing or evaluating architecture styles.
The English summary is being prepared. Showing the original description.