The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill acts as a scaffolding tool that ingests natural language descriptions and user answers to generate project artifacts, creating a potential vector for instructions to be embedded in produced files.
Ingestion points: The initial project description provided in the /project-starter command and answers provided to the question tool, as well as external content retrieved via context7_query-docs.
Boundary markers: There are no specific instructions or delimiters provided to the agent to treat user input or tool-provided documentation as data rather than instructions, potentially allowing embedded malicious prompts to influence the agent during the generation phase.
Capability inventory: The skill possesses file-writing capabilities to create documentation in docs/tech-decisions/ and project manifests like package.json, .env.example, and CI/CD workflow files (e.g., .github/workflows), which could be manipulated to include malicious code or configurations.
Sanitization: The instructions do not define any sanitization, filtering, or validation steps for user-supplied strings before they are used to populate project templates.

project-starter