plaid

Warn

Audited by Gen Agent Trust Hub on Apr 11, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the plaid-cli utility from a third-party GitHub repository (github.com/jverdi/plaid-cli@0.0.2) using the go install command.
  • [COMMAND_EXECUTION]: Executes the plaid-cli binary and various shell utilities (jq, sort, comm, mv) to query financial data and manage local state files.
  • [DATA_EXFILTRATION]: Accesses and processes highly sensitive financial data, including bank balances and transaction histories. The skill also manages Plaid access tokens and secrets, and it includes a specific warning about the tokens command to prevent unintentional exposure of authentication material.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted transaction data from external financial institutions. * Ingestion points: Data retrieved from plaid-cli transactions. * Boundary markers: Not present. The skill does not instruct the agent to use delimiters or ignore instructions embedded within the transaction content. * Capability inventory: Subprocess execution of the plaid-cli tool and shell commands. * Sanitization: None identified. The transaction data is processed directly through command-line pipelines.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 11, 2026, 01:24 AM