The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user data and project-specific context to create a self-contained prompt for a 'deep researcher' agent.
Ingestion points: The skill pulls context from conversation history, project files, and user-provided 'vague research needs' to fill its template.
Boundary markers: The resulting prompt lacks explicit instructions or delimiters to prevent the downstream agent from following malicious instructions that might be embedded in the user-supplied research topic or facts.
Capability inventory: The generated output is intended for execution via external research services (specifically mentioned as DeepAPI), which perform network-based searches and data retrieval.
Sanitization: There are no mechanisms described for sanitizing, escaping, or validating the input data before it is formatted into the final research brief.

research-prompt