java-classpath-search
Pass
Audited by Gen Agent Trust Hub on Mar 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill's instructions for searching the index (Step 3) and checking build output (Step 4) involve constructing shell commands like
grepandfindusing the$ARGUMENTSvariable. If the AI agent interpolates raw user input into these shell commands without escaping special characters, it could lead to the execution of unintended commands on the local system. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the following surface:
- Ingestion points: User-provided class names or patterns supplied via
$ARGUMENTS(SKILL.md). - Boundary markers: Absent; the instructions show direct interpolation of search patterns into shell commands without delimiters or warnings to ignore embedded control characters.
- Capability inventory: The skill uses
bash,grep,find, and a custom shell script (build-class-index.sh) which usesxargsandsh -cto execute commands. - Sanitization: Absent; the skill lacks explicit instructions or logic to sanitize, escape, or validate the user input before it is used to build shell commands.
Audit Metadata