ado-timesheet-report

Warn

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute system-level commands (PowerShell on Windows or the date command on Linux/macOS) to calculate and validate dates.
  • Evidence: In Phase 3, the instructions require the agent to validate a user-provided 'Specific week' end date by interpolating it into a shell command (e.g., date -d '[USER_INPUT]' +%u).
  • Risk: The instructions do not mandate strict regex validation or sanitization of the user-provided date string before it is passed to the shell. This could allow a malicious user to provide a crafted string containing command separators (like ';' or '&') to execute arbitrary code on the host environment.
  • [PROMPT_INJECTION]: The skill processes untrusted content from an external system (Azure DevOps) which represents an attack surface for indirect prompt injection.
  • Ingestion points: The skill ingests work item titles and descriptions via the 'wit_my_work_items' MCP tool.
  • Boundary markers: No explicit delimiters or boundary markers are defined to isolate external work item content from the agent's instructions in the final report output.
  • Capability inventory: The agent has access to the Bash tool (shell access) and Azure DevOps identity management tools.
  • Sanitization: While the instructions specify stripping HTML tags and truncating descriptions to 100 characters, no sanitization or filtering is applied to work item titles.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 28, 2026, 12:48 PM