The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external transaction data and merchant histories which could contain malicious instructions embedded by an external party.
Ingestion points: Financial data is ingested via tools such as list_transactions, search_transactions, and get_merchant_history as described in SKILL.md.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat transaction descriptors or merchant names as untrusted data.
Capability inventory: The skill utilizes a suite of read-only financial tools including list_accounts, get_spending_summary, get_recurring_charges, and list_holdings.
Sanitization: No evidence of data sanitization or filtering of external transaction strings was found in the instructions.

bankbridge