
from cli-agent-orchestrator954
Enables agents to report task outcomes and distill durable lessons into memory to improve performance across repeated workflow runs.
This skill transforms a standard agent workflow into a self-improving system. It allows agents to report the success or failure of specific work units, which are then processed by a 'retrospector' agent to identify patterns, traps, and efficient mappings, storing them as durable lessons for future sessions.
Use this skill in complex, iterative workflows that run repeatedly over similar work items (e.g., large-scale code migrations, data ETL conversions) where learning from previous failures is critical to efficiency.
Designed for orchestrator-style agent frameworks that support memory injection and multi-agent handoffs (e.g., custom CLI agent orchestrators, Claude Code, Codex).
This skill has not been reviewed by our automated audit pipeline yet.