idea-creator
Warn
Audited by Socket on May 17, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The core ideation and literature-review behavior is legitimate, and the OpenAI Codex path is plausible, but the skill is over-scoped: it has Bash(*) wildcard access, can orchestrate pilot experiments, mutates wiki/report files, uses opaque local helper scripts with unverifiable provenance, and includes a 'do it silently' directive. No confirmed credential theft or malicious payload is present, but the capability footprint is broader and riskier than a simple idea-generation skill.
Confidence: 84%Severity: 72%
Audit Metadata