mega-code-wisdom-gen

Pass

Audited by Gen Agent Trust Hub on Apr 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the uv tool to run Python modules (mega_code.client.*) located in a directory defined by a local configuration file (~/.local/share/mega-code/plugin-root). These commands handle session analysis, pipeline control, and retrieval of review instructions.
  • [PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface because it processes session data (untrusted input) through a pipeline and then instructs the agent to follow the resulting instructions exactly.
  • Ingestion points: Claude Code sessions processed by the extraction pipeline.
  • Boundary markers: None mentioned for the data analysis phase.
  • Capability inventory: The skill has permission to use Bash, Read, Write, and Edit tools, which could be leveraged if malicious instructions were successfully injected into the session logs.
  • Sanitization: The instructions do not specify any sanitization or validation steps for the session data being processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 10, 2026, 04:11 AM