skills/brycewang-stanford/awesome-agent-skills-for-empirical-research/marginaleffects/Gen Agent Trust Hub
marginaleffects
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a documentation assistant for the 'marginaleffects' statistical package. It provides instructions for an agent to answer user queries by reading local documentation files (Markdown and QMD files) related to the package's functionality and a companion book.
- [EXTERNAL_DOWNLOADS]: The skill references several external URLs (marginaleffects.com, routledge.com, and github.com/vincentarelbundock/marginaleffects). These are standard project homepages, publisher sites, and source code repositories. No automatic downloads or remote code executions are initiated by the skill.
- [COMMAND_EXECUTION]: The 'allowed-tools' section correctly restricts the agent to 'Read', 'Grep', and 'Glob' tools, which are appropriate for searching and reading the provided documentation files. There are no instructions to execute arbitrary shell commands or system-level operations.
- [INDIRECT_PROMPT_INJECTION]: The skill provides a surface for processing user-supplied queries by mapping them to local documentation content.
- Ingestion points: User queries trigger searches through local documentation files (e.g.,
chapters/*.qmd,man/**/*.md). - Boundary markers: The skill instructs the agent to classify the request and read specific relevant source files to provide structured responses.
- Capability inventory: The skill uses file-read tools ('Read', 'Grep', 'Glob') to access local documentation.
- Sanitization: The skill relies on the agent's ability to interpret documentation conceptually and provide code examples based on the framework provided.
Audit Metadata