power-bi-build

Pass

Audited by Gen Agent Trust Hub on Aug 5, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes the pbi CLI tool (provided by the pbi-cli-tool package) to interact with a Power BI MCP server on localhost for operations such as measure creation and model management.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the pbi-cli-tool package via pipx and utilizes the openpyxl library to parse Excel files.
  • [DATA_EXFILTRATION]: The skill reads from local Power BI Project files (.pbip) and Excel workbooks to extract data and metadata. It also accesses files within the ${WORK_FOLDER_ROOT} environment. While it processes sensitive business information, communications are restricted to localhost and no external network exfiltration was detected.
  • [PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection due to its processing of untrusted external data.
  • Ingestion points: Ingests analytical artifacts from the current session and reads headers, sheet names, and content from Excel workbooks.
  • Boundary markers: No explicit delimiters or instruction-ignoring markers are defined for the interpolation of external data into generated DAX code or JSON templates.
  • Capability inventory: The skill is capable of performing file system writes to create configuration files and executing shell commands via the pbi CLI.
  • Sanitization: The skill lacks explicit sanitization routines to validate or escape data retrieved from external Excel files before it is used in command arguments or dashboard definitions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 5, 2026, 07:54 PM
Security Audit — agent-trust-hub — power-bi-build