The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to dynamically create session directories (mkdir -p). The input used to generate these directory names is sanitized via a regex filter (/[^a-z0-9]+/g) and character limit to mitigate command injection risks.
[PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection. It ingests user-provided topics and chains the output of a 'Guidance' agent to multiple 'Role Analysis' agents, then synthesizes their results. A malicious topic or a compromised intermediate agent output could theoretically inject instructions that influence the behavior of subsequent agents in the pipeline. The skill uses delimited context propagation to reduce accidental obedience, but the surface remains.

maestro-brainstorm