cc-launch-execution
Pass
Audited by Gen Agent Trust Hub on Aug 2, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill defines a mechanism for creating subagent workflows based on the active conversation's plan, which acts as a vector for indirect prompt injection. * Ingestion points: Data enters the subagent context from the active conversation's plan. * Boundary markers: The instructions lack requirements for delimiters or specific prompt shielding to prevent subagents from following instructions embedded within the user-provided plan. * Capability inventory: Orchestrated workers using the Sonnet model are assigned tasks involving file modification and code implementation. * Sanitization: There are no documented steps for escaping or validating the content of the plan before it is passed to worker calls.
- [COMMAND_EXECUTION]: The skill utilizes dynamic orchestration to launch and manage background execution flows. * Patterns: The instructions specify the use of agent() for worker initialization and pipeline() for task chaining, which constitutes dynamic generation and execution of agent instructions.
Audit Metadata