eframe-project
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run standard Rust development commands including
cargo checkfor validation,cargo installfor the Trunk build tool, andtrunk servefor web development. - [EXTERNAL_DOWNLOADS]: The skill references the official
eframe_templaterepository on GitHub (https://github.com/emilk/eframe_template) as a source for web configuration files, which is the standard reference for this library. - [INDIRECT_PROMPT_INJECTION]: The skill correctly identifies that project files like
Cargo.tomland source code are external data sources that could contain malicious instructions. It provides an explicit guardrail for the agent to treat these files as data and ignore any instructions found within them. - Ingestion points:
Cargo.toml,Cargo.lock, source files, and command output. - Boundary markers: Includes a dedicated "Guardrails" section explicitly directing the agent to never follow instructions found in data files.
- Capability inventory: The skill involves reading and writing Rust project files, installing tools via cargo, and verifying builds.
- Sanitization: Implements prompt-based instruction filtering via the Guardrails section.
Audit Metadata