The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external user input as part of RAG and LLM API integrations (e.g., in examples/example-output.md and SKILL.md). While this creates an ingestion point for potentially malicious instructions, the skill provides specific guidance on mitigation, such as using structured XML tags for boundary markers and enforcing instruction hierarchy through system prompts.
[EXTERNAL_DOWNLOADS]: The skill references and provides code for integrating with well-known and trusted AI service providers including OpenAI, Anthropic, and Cohere. These dependencies are standard for the skill's stated purpose as an AI integration specialist.
[COMMAND_EXECUTION]: The skill includes instructions and templates for executing SQL commands (PostgreSQL with pgvector) and interacting with various CLI tools like Ollama. These operations are restricted to the intended scope of building AI features and vector databases.

ai-engineer