What problem does it solve? After a debugging, review, or audit investigation, turning findings into actionable work often means either proposing a quick inline fix or writing a vague ticket. This Skill bridges that gap by harvesting confirmed findings into self-contained, problem-describing prompts that a downstream planning agent can execute, without prescribing the implementation. ## Core Features & Use Cases - Investigation Harvesting: Collects symptoms, root causes with file:line references, evidence, hypotheses, and open questions already present in the conversation. - Defect Clustering: Groups related defects into the smallest sufficient number of clusters with per-cluster rationale, merging related issues and splitting only genuinely independent ones. - Prompt Delegation: Hands each cluster to /prompt-architect, which composes and saves prompts under .agents/prompts/ and owns the launch confirmation for agent-quorum runs. - Use Case: After tracing three related test failures to a shared root cause in the provider layer, invoke this Skill to produce one clustered handoff prompt that a downstream agent uses to design a systemic, compatibility-preserving fix. ## Quick Start Ask the agent to run /solution-handoff after your investigation is complete so it clusters the confirmed defects and delegates handoff prompts to /prompt-architect.
