jdb-debugger

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses Bash scripts (jdb-launch.sh, jdb-attach.sh, jdb-diagnostics.sh) to execute the Java Debugger (jdb) and system utilities like nc and timeout for JVM connectivity and lifecycle management.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data that influences the execution of powerful debugger commands. \n
  • Ingestion points: CLI arguments in scripts/jdb-launch.sh, scripts/jdb-attach.sh, and scripts/jdb-breakpoints.sh (e.g., --mainclass, --bp, --cmd). \n
  • Boundary markers: None present in the scripts or prompt instructions to distinguish between debugger control commands and user data. \n
  • Capability inventory: The skill can execute arbitrary Java code via JDB's print and eval commands, and interact with the filesystem or network via the debugged JVM's context. \n
  • Sanitization: The scripts do not sanitize or validate user-provided class names or debugger commands before passing them to the JDB process.
  • [COMMAND_EXECUTION]: The jdb-breakpoints.sh script demonstrates dynamic execution patterns by generating temporary command files and piping them to the jdb process to automate complex debugging workflows.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 02:14 PM