The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a process for characterizing data sources such as files, databases, and APIs. Ingesting untrusted data from these sources while the agent has access to powerful tools like Bash introduces a vulnerability surface where instructions hidden in the data could be interpreted by the agent. \n- Ingestion points: Source inventory includes file drops, database queries, and API endpoints (SKILL.md). \n- Boundary markers: The instructions do not specify delimiters or warnings to ignore instructions embedded within the source data. \n- Capability inventory: The skill utilizes Read, Grep, Glob, and Bash tools to perform source analysis. \n- Sanitization: The workflow does not include specific steps for sanitizing, escaping, or validating the contents of the analyzed data sources.

source-analyzer