shinka-run

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill ingests data from local task files and user feedback to generate system prompts for evolutionary algorithms. 1. Ingestion points: File content in the task directory and user-provided feedback strings. 2. Boundary markers: No explicit delimiters are used to isolate instructions within the ingested content. 3. Capability inventory: The shinka_run CLI tool performs code evaluation, and the skill uses ls for directory inspection. 4. Sanitization: No explicit sanitization or filtering of input data is performed.
  • [COMMAND_EXECUTION]: The skill uses the shinka_run CLI and ls command. These are standard for the skill's purpose and are documented by the vendor.
  • [SAFE]: All external URLs and the CLI tool are official resources from the author, SakanaAI, representing normal vendor functionality.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 10:04 AM