The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the claude CLI command with the --bg flag to spawn and manage background agent processes. This is the primary intended function of the skill.
[DATA_EXPOSURE]: The instructions contain a specific safety requirement for the agent to redact sensitive information, such as API keys, passwords, and PII, from the summary before it is used as a prompt for the new session.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the transfer of conversation context to a new agent. While this creates a surface for indirect prompt injection, the explicit instruction to summarize and redact content provides a natural mitigation layer.

claude-handoff