engineering-frontend-developer
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: Includes a utility script 'scripts/check_bundle.sh' which uses standard shell commands like 'find', 'stat', and 'awk' to traverse build directories and analyze file metadata. This script is a legitimate vendor resource provided by the author for performance optimization tasks.
- [PROMPT_INJECTION]: The codebase orientation and debugging instructions direct the agent to ingest and analyze files from external projects (e.g., component source, package.json, and documentation). This creates a surface for indirect prompt injection if those external files contain malicious instructions.
- Ingestion points: Orientation phase where the agent reads the project's component tree and metadata.
- Boundary markers: None identified; the instructions do not specify ignoring instructions or using delimiters for untrusted file content.
- Capability inventory: The agent has the capability to execute local scripts (bash) and package manager commands (npm/npx) in the environment.
- Sanitization: None; the skill does not define validation or filtering of content read from the target project files.
Audit Metadata