The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its reliance on specifications defined in external documents.
Ingestion points: The agent reads initialization requirements from specs/技术栈.md, specs/项目结构.md, and specs/开发规范.md (referenced in SKILL.md).
Boundary markers: There are no markers or instructions used to separate the specification data from agent instructions, making the agent susceptible to embedded prompts within those files.
Capability inventory: The skill has the ability to execute shell commands (e.g., git init, mkdir, npm init) and modify the local filesystem (SKILL.md).
Sanitization: The skill does not perform validation or sanitization on the file content before using it to define command execution paths.
[COMMAND_EXECUTION]: The skill performs project setup using standard shell utilities.
Evidence: Workflow steps in SKILL.md explicitly call for the execution of git init, mkdir -p, and technical stack initialization commands such as npm init or go mod init.

project-initialization