The Agent Skills Directory

[SAFE]: No security issues were detected. The skill follows best practices for agent orchestration, including the use of subagent isolation to prevent context pollution and mandatory user steering gates to ensure human alignment before executing creation or modification tasks. The workflow is well-documented and does not include any suspicious command execution, data exfiltration, or obfuscation techniques.
[INDIRECT_PROMPT_INJECTION]: The workflow involves a Coordinator agent ingesting research reports from subagents. While this is a potential ingestion surface for untrusted data from external sources, the skill mitigates the risk through a mandatory human-in-the-loop steering gate and context isolation. This ensures all synthesized data is reviewed by the user before being used in the development phase.

double-diamond