go
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The instructions explicitly direct the agent to access sensitive file paths (such as .env files) or use a secrets manager to retrieve credentials for testing. This pattern of accessing sensitive configuration files increases the risk of data exposure.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes an indirect prompt injection surface by requiring the agent to ingest data from an external, potentially untrusted source (browser UI content, console logs, and network responses) and then perform code modifications based on that data.
- Ingestion points: Browser console output, network request failures (4xx/5xx), and rendered UI content from a running application.
- Boundary markers: None present in the instructions to separate untrusted UI data from agent commands.
- Capability inventory: The agent is authorized to modify the codebase ('fix and verify again') based on the interpreted browser data.
- Sanitization: No sanitization or validation of the browser-derived data is specified before it influences the agent's actions.
Audit Metadata