The Agent Skills Directory

[Prompt Injection] (LOW): The skill is designed to ingest and process untrusted data from multiple external sources, creating an attack surface for indirect prompt injection.
Ingestion points: Processes content from ~~chat, ~~email, ~~cloud storage, and ~~project tracker as seen in the 'The Goal' and 'Synthesis Workflow' sections.
Boundary markers: No specific delimiters (e.g., XML tags or triple quotes) or 'ignore' instructions are mandated to separate untrusted content from the agent's instructions.
Capability inventory: The skill is limited to generating narrative summaries for user display; it lacks direct file-system access, command execution, or network capabilities.
Sanitization: The instructions provide no guidance on sanitizing or filtering malicious instructions embedded within the source data.
[No Code] (INFO): This skill contains only documentation and natural language instructions. There are no executable scripts, binaries, or configuration files that could facilitate traditional RCE or persistence.

knowledge-synthesis