The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, which creates a surface for indirect prompt injection attacks.
Ingestion points: The skill reads repository content, AI-generated changes, benchmark datasets, and distributed trace logs using tools like read_file and search_code.
Boundary markers: There are no instructions defining the use of delimiters or 'ignore embedded instructions' markers when the agent processes external file content.
Capability inventory: The skill utilizes powerful tools including execute_command, write_file, edit_file, and create_file.
Sanitization: The instructions do not specify validation or sanitization requirements for the content retrieved from external sources before it is used in logic or further commands.

agent-quality-gate