The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external financial data, which creates a potential vector for malicious instructions embedded in the data.
Ingestion points: Phase 0 explicitly requests data from external sources such as 'ledger records', 'order books', and 'platform exports' (CSV/Excel files).
Boundary markers: The instructions do not define specific delimiters or security framing to prevent the agent from being influenced by instructions that might be hidden within these external data files.
Capability inventory: The skill reads local files including config.yaml, historical memory paths, and instruction documents from the references/ directory.
Sanitization: There is no mention of sanitizing or validating the input data before processing it for analysis.

aikey-revenue