alibabacloud-writing-plans

Pass

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-influenced design specifications (design.md) and interpolates this content into instructions for downstream code generation skills. While this represents a standard data ingestion surface, it is a necessary part of the skill's primary function.
  • Ingestion points: The agent reads infrastructure specifications from .aliyun-ai-ops-spec/{name}/designs/design.md.
  • Boundary markers: No explicit delimiters or instructions to disregard embedded commands are present when passing data to the terraform-codegen skill.
  • Capability inventory: The skill utilizes tool invocation (delegation to other skills) and local file writing.
  • Sanitization: There is no explicit sanitization step for the design content before it is processed.
  • [COMMAND_EXECUTION]: The skill generates shell scripts (commands.sh) containing Alibaba Cloud CLI commands (aliyun). These scripts are intended for manual or automated execution in later stages. The skill instructions specifically exclude direct execution of these scripts during the planning phase.
  • [DATA_EXFILTRATION]: The skill interacts with state and configuration metadata within the .aliyun-ai-ops-spec/ directory. These operations are scoped to the local environment and are used to track the progress of infrastructure tasks (e.g., status updates in status.json).
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 12, 2026, 01:50 PM
Security Audit — agent-trust-hub — alibabacloud-writing-plans