curl-command-generator
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides templates for generating shell scripts (e.g., api-commands.sh) and Makefiles that execute curl commands. The generated Bash script includes a dispatcher pattern ("$@") that executes user-supplied arguments as commands, which is a common and documented feature for this type of command-line wrapper utility.
- [INDIRECT_PROMPT_INJECTION]: The skill contains logic to scan project source code to identify API routes and extract metadata for command generation. This creates an attack surface where the agent processes external file content.
- Ingestion points: API route definitions and source files (Express, Next.js, Fastify) in the source directory.
- Boundary markers: The code does not implement specific boundary markers or delimiters to separate content extracted from source files in the generated output.
- Capability inventory: The skill uses the fs.writeFileSync method in its CLI script to save generated commands to the file system.
- Sanitization: The provided scripts perform basic string replacement for path parameters (e.g., converting :id to {id}) but do not include comprehensive shell escaping or sanitization for all extracted metadata.
Audit Metadata