The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates task handoff by instructing the agent to execute the claude CLI command with the --bg flag. This is the core intended functionality of the skill and leverages the platform's native tooling.
[INDIRECT_PROMPT_INJECTION]: The skill uses conversation history and user-supplied arguments to seed a new agent session, which creates an inherent surface for indirect prompt injection.
Ingestion points: Untrusted data enters the context from the current conversation and user arguments via the instructions in SKILL.md.
Boundary markers: The summary content is wrapped in double quotes in the command template, but no explicit markers are defined to isolate the summary content from the rest of the new agent's prompt.
Capability inventory: The skill possesses the capability to execute shell commands (specifically the claude binary) as defined in SKILL.md.
Sanitization: The instructions explicitly command the agent to redact sensitive information such as API keys, passwords, and PII, which serves as a mitigation for data exposure, although it does not provide specific instructions for escaping shell-sensitive characters.

claude-handoff