mercury-autocategorize-transactions
Warn
Audited by Snyk on Jun 13, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (medium risk: 0.65). Outsider free text can enter the LLM context via Step 2/Step 4 when the skill fetches Mercury transaction fields (e.g.,
counterpartyNameanddescription) frommercury transactions list --format jsonl, since those strings are authored by third parties (vendors/transaction counterparties) and then rendered into the review table that the agent/LLM uses.
MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).
- Direct money access detected (high risk: 1.00). The skill explicitly calls Mercury banking CLI endpoints to read and write account data and is designed to operate on bank transactions. It fetches transactions and categories via
mercury transactions list/mercury categories listand performs writes withmercury transactions update --transaction-id <txn.id> --category-id <suggestion.categoryId>. This is a specific banking API integration that performs updates to bank transaction records (direct financial/banking operation), even though writes require user confirmation.
Issues (2)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
W009
MEDIUMDirect money access capability detected (payment gateways, crypto, banking).
Audit Metadata