YunxiaoPMapp

Fail

Audited by Gen Agent Trust Hub on Jul 24, 2026

Risk Level: HIGHCREDENTIALS_UNSAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The automation script scripts/live_create_fast.py utilizes the browser_cookie3 library to programmatically extract authentication cookies from the user's local Chrome browser profile for the .aliyun.com domain. This allows the skill to obtain and use the user's active session credentials for Alibaba Cloud DevOps without standard authentication prompts.\n- [COMMAND_EXECUTION]: The skill relies on the execution of local Python scripts (scripts/live_create_fast.py and scripts/workday_hours.py) to perform its core functions, including network communication with the Yunxiao API and complex workday calculations.\n- [DATA_EXFILTRATION]: The skill extracts sensitive session data from the local browser and transmits it to the well-known service domain devops.aliyun.com to facilitate its automated tasks.\n- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes data from external sources on the Yunxiao platform. 1. Ingestion points: Requirement descriptions and work item content retrieved from devops.aliyun.com. 2. Boundary markers: None identified; instructions and data are processed together in text fields. 3. Capability inventory: Subprocess execution for scripts that can create and modify project work items, change statuses, and update assignees. 4. Sanitization: No input validation or escaping is applied to the ingested content before it is processed by the agent or included in API payloads.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jul 24, 2026, 09:34 AM