helpmetest-debugger
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The instructions focus exclusively on the technical debugging workflow. No patterns suggesting attempts to bypass safety filters, extract system prompts, or override agent constraints were found.
- [DATA_EXFILTRATION]: No evidence of hardcoded credentials or unauthorized data transmission. Access to authentication states and URLs is strictly for the purpose of test reproduction within the targeted application environment.
- [REMOTE_CODE_EXECUTION]: The skill does not perform external downloads of executable scripts. Interactions are handled through designated MCP tools (
mcp__helpmetest-*). - [COMMAND_EXECUTION]: While the skill generates dynamic automation commands (Robot syntax), these are scoped to a browser automation environment via specific tools and do not pose a risk of arbitrary shell command execution.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests test data and feature descriptions which could theoretically contain instructions. However, the risk is minimal given the specialized nature of the interactive debugging toolset and the vendor-controlled environment.
Audit Metadata