The Agent Skills Directory

[Prompt Injection] (SAFE): The skill uses persona-based instructions and emphasis markers like 'CRITICAL', which are appropriate for its primary purpose and do not attempt to bypass safety guidelines.
[Indirect Prompt Injection] (SAFE): The skill implements a retrieval-augmented generation (RAG) workflow. 1. Ingestion points: User input in Step 1 and tool results in Step 5. 2. Boundary markers: No explicit delimiters are used for interpolating tool outputs into the final prompt. 3. Capability inventory: Access is restricted to specific read-only MCP persona-retrieval tools. 4. Sanitization: No specific sanitization or escaping of external data is documented beyond the synthesis process.

copywrite-experts