alicloud-network-cdn-test
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script 'tests/common/compile_skill_scripts.py' to verify the compilation of target skill scripts.
- [PROMPT_INJECTION]: The skill reads the 'SKILL.md' file of a target skill to dynamically determine which API to execute, creating a surface for indirect prompt injection. Ingestion points: Target skill 'SKILL.md' file. Boundary markers: None present to distinguish instructions from data. Capability inventory: Local Python script execution and Alibaba Cloud CDN API queries. Sanitization: None provided; the skill relies on the agent to identify and select read-only, lowest-risk API calls.
Audit Metadata