loop
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEPERSISTENCECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PERSISTENCE]: The skill utilizes the 'CronCreate' tool to schedule recurring autonomous tasks that execute instructions over time.- [COMMAND_EXECUTION]: The skill instructs the agent to execute multiple shell commands, including 'git' branch management, commits, and a Python script ('run_experiment.py').- [INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection. • Ingestion points: Reads configuration, strategy, and results from '.autoresearch/' directory files. • Boundary markers: Lacks delimiters or instructions to ignore embedded commands within the ingested data. • Capability inventory: Includes file modification, git operations, and script execution. • Sanitization: Does not sanitize or filter the content of history files before using them to drive its autonomous code modification strategy.
Audit Metadata