The Agent Skills Directory

[SAFE]: No malicious patterns, data exfiltration, or obfuscation were detected during the analysis. The skill focuses entirely on efficient orchestration and task management.
[COMMAND_EXECUTION]: The skill provides a template for using the pi CLI tool to launch worker sessions. The instructions emphasize security best practices, such as using a narrow tool allowlist (e.g., --tools read,grep,find,ls for read-only workers) and assigning security-sensitive validation tasks to specific reasoning-capable models.
[PROMPT_INJECTION]: The instructions are procedural and technical, focusing on model routing and orchestration logic. There are no attempts to bypass safety filters, extract system prompts, or override agent constraints.
[DATA_EXFILTRATION]: No network exfiltration or unauthorized file access patterns were identified. The orchestration steps focus on maintaining bounded contexts and reducing token usage.

model-routing