The Agent Skills Directory

[SAFE]: The skill implements a structured framework for decision analysis and does not contain any malicious instructions or patterns. It follows best practices for interactive agent tasks.
[COMMAND_EXECUTION]: While the skill utilizes tools like Read, Grep, and Glob, these are used within the context of the agent's environment to retrieve framework information and project context. No arbitrary command execution or privilege escalation vectors were found.
[DATA_EXFILTRATION]: The skill does not perform any network operations. The Write tool is used solely for saving decision journals to a local directory (docs/decisions/) within the user's project, which is a transparent and intended behavior of the tool.
[EXTERNAL_DOWNLOADS]: No external dependencies, remote scripts, or unverified package installations were detected. The use of the mcp__solograph__kb_search tool is a configuration for an existing environment capability and does not involve downloading external code.

solo-stream