client-package-local
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from multiple local file types (.md, .csv, .json) and processes their content to generate summaries and Google Sheets. • Ingestion points: Files located in the
clients/<client_name>/subdirectories. • Boundary markers: Absent. The instructions do not provide delimiters or warnings to ignore instructions embedded within the source files. • Capability inventory: The skill has permissions to read/write files, create directories, and perform network operations via the Google Sheets API. • Sanitization: No sanitization or filtering is applied to the content read from the source files before it is processed by the agent. - [DATA_EXFILTRATION]: The skill reads lead lists and strategy documents from the local filesystem and uploads them to Google Sheets via an MCP server. This facilitates the transfer of local data to an external cloud service.
- [COMMAND_EXECUTION]: The skill executes shell commands (
mkdir -p) to create the output directory structure for the delivery package, using parameters derived from user input.
Audit Metadata