handoff

Summary

A meta-skill that compacts the current conversation into a handoff document for another agent to continue work. The skill summarizes context, redacts sensitive data, and saves to the OS temporary directory with a "suggested skills" section. It is designed to bridge agent-to-agent handoffs by avoiding duplication of existing artifacts and providing clear continuation context.

Detected Capabilities

Trigger Keywords

Phrases that agents use to match this skill to user intent.

Risk Signals

Use Cases

• Handing off a partially completed task to another agent in a multi-agent workflow
• Summarizing conversation history when session context is lost or reset
• Creating continuity artifacts before an agent session expires
• Guiding a fresh agent on which domain-specific skills to activate for the next phase

Quality Notes

• Clear, concise instructions with explicit scope (handoff, not artifact creation)
• Strong security posture: explicitly requires redaction of API keys, passwords, and PII
• Good guidance on avoiding duplication by referencing existing artifacts instead of copying
• Scope boundary is well-defined (temporary directory, not workspace)
• disable-model-invocation: true is appropriate for a deterministic task
• Argument hint ('What will the next session be used for?') provides good context for user intent
• Missing explicit examples of what a handoff document should contain (structure, sections, format)
• No guidance on how to identify or reference existing artifacts (specs, plans, ADRs, issues, commits, diffs)
• Does not specify what 'suggested skills section' format should look like or how to name skills