tres-request-skill-update
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- Command Execution for Identity Collection: The skill uses
git configand system calls to retrieve the user's name and email address. While used to attribute feedback, this involves executing shell commands to access local configuration data. - Feedback Data Ingestion (Indirect Prompt Injection): The skill processes user-provided feedback and conversation history. While it includes redaction steps for tokens and paths, this ingestion surface could be used to influence the agent's behavior if malicious instructions are embedded in the feedback text.
- External Data Transmission: The skill is designed to send structured feedback and recent conversation history (up to 30 exchanges) to an external TRES MCP tool. This constitutes a data transfer that includes context from the user's session.
- Privacy and Redaction Measures: The skill implements a redaction process to remove sensitive patterns such as Bearer tokens, hex strings, and specific file paths before submission. This is a positive security practice, though users should confirm that the 'Preview' step accurately reflects what will be sent.
Audit Metadata