The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the exploration of untrusted structured data from various file formats (JSON, YAML, TOML, XML, CSV, HCL, INI), creating a surface for potential indirect prompt injection if those files contain malicious instructions.
Ingestion points: The skill reads file contents using cat and pipes them to the dasel CLI for inspection (e.g., cat config.yaml | dasel ...).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore embedded instructions within the data files.
Capability inventory: The skill uses shell commands and the dasel tool to query and extract data from files.
Sanitization: The skill does not provide methods for sanitizing the input data before processing, which is standard for data exploration tools.

data-exploration