java-route-mapper
Warn
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The
SKILL.mdfile uses forceful language and 'CRITICAL' markers to override agent behavior and control the output flow.\n- [PROMPT_INJECTION]: The skill implements a custom validation logic that directs the agent to 'stop current output' and perform re-analysis, which is a technique for hijacking the AI's standard operational loop.\n- [COMMAND_EXECUTION]: The skill utilizes external MCP tools for decompiling Java class files based on user-supplied paths, such asmcp__java-decompile-mcp__decompile_directory.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. 1. Ingestion points: Java source code and XML configurations. 2. Boundary markers: None present. 3. Capability inventory: Decompilation tools and file-writing operations. 4. Sanitization: Absent.
Audit Metadata