Review Agent Memory

Review Agents — Memory and State

Overview

Scans agent memory and state-store configuration (component YAMLs + in-code wiring) for correctness: actorStateStore: "true" on the workflow state store, appropriate memory class selection, secret management on LLM keys, and no component collisions between conversation memory and workflow state. Read-only: this skill never modifies source files.

Execution Order

You MUST follow these phases in strict order.

1. Resolve scope — Read ../shared/review-scope-prompt.md and follow it to set scope_root.
2. Detect target — Read ../shared/review-detect-target-agent.md and follow it to produce language, flavor, agent_files, and component_files.
3. Load checklist — Based on language, read exactly one of:
   • python → ../shared/review-agent-memory-python.md
   • dotnet → ../shared/review-agent-memory-dotnet.md
4. Scan — Apply every rule from the loaded checklist to agent files and component YAMLs. Cross-reference: a rule about a component is only valid if the component name is referenced in code (or conventionally expected, like agent-workflow).
5. Report — Format findings using ../shared/review-report-format.md. Group by severity, then rule id, then file path.
6. Show final message — Emit the report from step 5 with a ## Next steps block.

Prerequisites

• Read access to the project directory.
• No build, compile, or run step required.

Allowed tools

Read, Grep, Glob only.

Rules

• Python rules — DAG-MEM-001 … DAG-MEM-008
• .NET rules — DAG-MEM-001 … DAG-MEM-007

Show final message

The last thing you emit MUST be the report from step 5. ## Next steps should suggest:

• Run review-agent-tools next if not yet run.
• Run review-agent-orchestration for multi-agent projects.
• Run review-agent-observability for native dapr-agents projects.
• Re-run this skill after fixing any critical findings.

Do not add any text after the report.