The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and transform untrusted structured data (JSON, YAML, CSV, XML, etc.) from external files.
Ingestion points: The skill reads data from various files using cat and pipes it to the dasel utility as shown in SKILL.md and references/transformation-patterns.md.
Boundary markers: There are no explicit instructions to the agent to treat data inside these files as untrusted or to ignore any embedded instructions that might be contained within values.
Capability inventory: The skill uses dasel for transformation and shell redirection (>) and mv for file writing. It does not perform network operations or execute arbitrary scripts.
Sanitization: The skill does not implement or describe sanitization techniques for the data being processed, relying on dasel for structure parsing.

data-transformation