T05 · Unauthorized Access and Privilege Escalation
- Location
SKILL.md:7- Finding
Overly Broad Local Code Execution and Filesystem Permissions
- Content
View full analysis
- Remediation
View remediation
Security audit
Security checks for vulnerabilities and agentic risk
The skill appears aimed at legitimate creative media generation, but it asks for broad local execution and file access while the referenced implementation files are missing from the package.
Review this carefully before installing. Use it only in a sandboxed project directory with non-sensitive inputs and provider API keys you are comfortable using for media generation. Avoid giving it access to private repositories, credentials, or broad home-directory files until the missing scripts/templates are included and command permissions are narrowed.
SKILL.md:7Overly Broad Local Code Execution and Filesystem Permissions
The skill description is framed broadly enough that an agent could invoke it for many loosely related creative tasks without strong activation boundaries or explicit user confirmation. Because the skill can make network requests, generate third-party content, and write files, overbroad routing increases the chance of unintended execution and external data disclosure during ordinary prompts.
The documentation describes capabilities and provider routing but does not prominently warn that prompts, uploaded media, and possibly project content are transmitted to third-party APIs and that output files and task-tracking artifacts are written locally. This can cause users or higher-level agents to disclose sensitive content or modify the filesystem without informed consent.