brainstorming

SUSPICIOUS: the skill’s core brainstorming purpose is legitimate, but it expands its footprint with mandatory external MCP usage, transitive skill invocation, and write/commit behavior after inspecting untrusted content. The install provenance cited is mostly official, so this is not confirmed malware; the main concerns are scope expansion, prompt-injection exposure, and downstream trust chaining.

The code is a legitimate server cleanup script with no clear malicious behavior. It has a security risk because the /tmp-only deletion check is based on a string prefix rather than a canonical path, allowing traversal-style destructive deletion if an attacker can control SESSION_DIR. The unvalidated PID file also permits unintended process signaling. Input validation, canonicalization, and strict PID validation are recommended.