UltraPlan Autonomous Planner
Warn
Audited by Gen Agent Trust Hub on Mar 26, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The
scripts/generate-plan.shscript executes Python logic using thepython3 -cflag with shell variables interpolated directly into the command string. This pattern is vulnerable to command and code injection if variables like the configuration file path or default values are manipulated to contain malicious escape characters. - [PROMPT_INJECTION]: The skill ingests untrusted input from
resources/plan-config.jsonand incorporates it into generated strategic plans without using sanitization, escaping, or boundary markers. This establishes an indirect prompt injection surface where malicious instructions embedded in the configuration data could influence the agent's behavior when it later processes the generated plan. - [DATA_EXFILTRATION]: Documentation in
SKILL.mdhighlights features for integrating with external CRM systems and configuring webhooks to remote endpoints. These capabilities, while described as business integrations, represent a potential vector for data exfiltration if the agent is instructed to send sensitive data to attacker-controlled destinations.
Audit Metadata