Agent
Use this agent when reviewing service topology, scalability, high availability, network design, Terraform component structure, or Atmos integration patterns. Dispatch for plan reviews or infrastructure code reviews.
Agent
Use this agent when reviewing backend application code (Java, Kotlin, Python, Node/TypeScript, Go). Detects the stack from repo markers (pom.xml, build.gradle, pyproject.toml, package.json, go.mod), loads relevant agnostic and framework-specific skills, and dispatches cross-cutting concerns to specialist agents. Only…
Agent
Grade PM10 — does this serve actual business goals, or is it engineering-driven scope creep? Dispatched by /plan-review, /research, and standalone PM workflows.
Agent
Use this agent for a holistic infra-code review using the AWS Well-Architected Framework across all 6 pillars: Operational Excellence, Security, Reliability, Performance Efficiency, Cost Optimization, and Sustainability. Dispatch for architecture decision reviews, production readiness gates, or when evaluating overall…
Agent
Use this agent when evaluating plan clarity, actionability, software architecture quality, or developer experience. Always dispatch for plans with stories.
Agent
Use this agent when reviewing cost optimization — resource right-sizing, environment tiering, NAT gateway patterns, expensive resource toggles, and FinOps best practices. Dispatch for plan reviews or infrastructure code reviews.
Agent
Grade PM11 — does the research synthesis honor the framing brief's PF7 must-cite voices and PF8 source-type coverage matrix? Each PF7 voice must appear with a direct quote in the body; each PF8 required source type must have at least one direct quote in the body. Research-Review only. Dispatched by /research Phase 2…
Agent
Grade PM6 — how does this position relative to alternatives (buy-vs-build, existing tools, competitor offerings)? Dispatched by /plan-review, /research, and standalone PM workflows.
Agent
Use this agent when reviewing Kubernetes manifests, Helm charts, or Kustomize overlays for security, cost, or operational concerns. Dispatch for K8s-specific code reviews when K8s manifests are detected. Only activate when there is clear evidence of Kubernetes usage — if ambiguous, ask the user.
Agent
Grade PM4 — is effort-vs-impact considered, dependencies mapped, sequencing justified? Dispatched by /plan-review, /research, and standalone PM workflows.
Agent
Grade PM2 — does the proposed approach actually solve the stated problem, or is it a solution looking for a problem? Dispatched by /plan-review, /research, and standalone PM workflows.
Agent
Use to frame a research topic BEFORE parallel research streams run. Produces an 8-section structured brief (PF1-PF8) that shapes downstream agent prompts — stakeholders, decisions, success criteria, must-cite voices, source-type coverage matrix. Dispatched by /research Phase 2 framing.
Agent
Use AFTER research synthesis to produce the narrative portion of Research-Review (User Impact Analysis, Scope Recommendation, Prioritization Framework, Stakeholder Summary). The graded PM1-PM11 scorecard is handled by separate dim subagents, not by this one. Dispatched by /research Phase 2 review.
Agent
Grade PM9 — what happens if this is the wrong bet? How hard is it to change course? Dispatched by /plan-review, /research, and standalone PM workflows.
Agent
Grade PM3 — is this an MVP or a kitchen sink? What can be cut without losing core value? DISTINCT from the PRD-Review dim 2 scope-discipline prompt (which grades a PRD's Non-Goals section). Dispatched by /plan-review, /research, and standalone PM workflows.