qa-performance-reliability
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external performance reports which may contain untrusted content.\n
- Ingestion points: Files like lighthouse-report.html and k6-summary.json are read into the agent's context.\n
- Boundary markers: The instructions do not specify any delimiters or safety prompts to prevent the agent from executing instructions found within these reports.\n
- Capability inventory: The skill has access to Bash, Read, Grep, and Glob tools, which could be misused if an injection occurs.\n
- Sanitization: There is no mention of sanitizing or validating the input from these reports before processing.\n- [COMMAND_EXECUTION]: Executes shell commands via Bash to perform performance audits and smoke tests.\n
- Evidence: Uses lhci autorun and k6 run commands.\n- [DATA_EXPOSURE_AND_EXFILTRATION]: Accesses internal configuration paths within the agent's home directory.\n
- Evidence: Interacts with files in ~/.claude/orchestrator-runtime/qa/schemas/ and ~/.claude/orchestrator-runtime/qa/presets/. These directories house runtime and policy configurations specific to the sensliao orchestration framework.
Audit Metadata