weekly-report-to-annual
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFECREDENTIALS_UNSAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- CREDENTIALS_UNSAFE (LOW): The skill's
fetch_emailsscript (defined inSKILL.md) requires an email address and password as arguments. Passing secrets as command-line arguments is a known security risk as they can be visible in process monitoring tools. - PROMPT_INJECTION (LOW): The skill is vulnerable to Indirect Prompt Injection (Category 8) because it processes external email data.
- Ingestion points:
fetch_emails.py(referenced inSKILL.md) retrieves content from external emails. - Boundary markers: Absent; the template does not specify delimiters or warnings to separate untrusted email content from agent instructions.
- Capability inventory:
save_report.pyallows writing to the local filesystem. - Sanitization: No sanitization logic is present in the provided scripts to filter or escape malicious instructions within emails.
- COMMAND_EXECUTION (LOW): The
save_report.pyscript allows the agent to write arbitrary content to user-defined paths usingPath.expanduser(). While restricted by system-level user permissions, this capability allows for arbitrary file creation or overwriting if the agent is misled.
Audit Metadata