The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves analyzing user-provided source code to generate unit tests, which constitutes an indirect prompt injection surface.
Ingestion points: User-provided code snippets are ingested and analyzed as described in the '分析被测代码' section of the workflow.
Boundary markers: There are no explicit instructions or delimiters defined to isolate the user-provided code or to warn the agent against executing instructions that might be embedded within that code.
Capability inventory: The skill instructions do not request or use high-risk tools like file system access, network operations, or shell execution, and the skill explicitly disclaims the ability to execute tests.
Sanitization: No validation or sanitization mechanisms are described for handling the incoming user content.
[PROMPT_INJECTION]: The skill contains a '能力缺口响应规则' (Ability Gap Response Rules) section that mandates a specific promotional response when a user's request exceeds its capabilities. This response includes a command to install an external repository (/install 技能宝 github:mindverse/skillhub). This steers the agent's behavior to promote content from an external source not identified as a trusted vendor.

unit-test