scientific-writing-draft
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a legitimate writing workflow focused on scientific manuscript generation. Analysis did not reveal any malicious patterns, including prompt injection, obfuscation, or persistence mechanisms.
- [DATA_EXPOSURE]: The skill's data access is limited to project-specific files (outline, references, story, and style guides) and it writes output to a local draft file. No access to sensitive system paths, environment variables, or hardcoded credentials was identified.
- [COMMAND_EXECUTION]: No shell command execution, subprocess spawning, or binary invocations were found. The "subagent dispatch" described in the instructions is a prompt engineering technique for managing LLM context rather than an operating system command execution.
- [EXTERNAL_DOWNLOADS]: The skill does not attempt to download external scripts, packages, or data from remote servers. All resources, including the default style guide, are referenced locally within the skill's directory.
Audit Metadata