plaid

Warn

Audited by Gen Agent Trust Hub on Oct 26, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to download and install an external binary from a third-party GitHub repository (github.com/jverdi/plaid-cli@0.0.2) using the go install command.
  • [REMOTE_CODE_EXECUTION]: The skill uses go install to fetch and execute code from an unverifiable external source at runtime.
  • [DATA_EXFILTRATION]: The skill accesses and manages highly sensitive financial data, including PLAID_CLIENT_ID, PLAID_SECRET, and access tokens. It specifically references sensitive local files and directories such as ~/.plaid-cli/config.toml and ~/.plaid-cli where these credentials and tokens are stored.
  • [COMMAND_EXECUTION]: The skill relies on extensive shell command execution, including piping output through utilities like jq, sort, comm, and mv for processing financial data.
  • [PERSISTENCE]: The skill suggests establishing persistence on the host system by using cron to schedule recurring transaction monitoring scripts.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill retrieves transaction data from external banking institutions via the Plaid API (plaid-cli transactions).
  • Boundary markers: There are no instructions provided to the agent to treat transaction data as untrusted or to ignore instructions that might be embedded in transaction names or descriptions.
  • Capability inventory: The skill has the capability to execute shell commands and file system operations based on the processed transaction data.
  • Sanitization: The skill does not implement sanitization or validation of the data retrieved from the external API before processing it with shell utilities.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Oct 26, 2026, 09:37 AM