memory-leak-debugging
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data by interacting with web pages and capturing heap snapshots, which could contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: Browser data from pages (via
navigate_page,fill) and process memory (viatake_heapsnapshot) are ingested into the agent's context during debugging. - Boundary markers: The instructions do not define boundary markers or delimiters to separate the skill's logic from the data being analyzed.
- Capability inventory: The skill utilizes the
chrome-devtools-mcpserver tools which have capabilities for browser interaction and system-level heap snapshot management. - Sanitization: There is no evidence of sanitization or filtering applied to the snapshot content or page data before processing by the agent.
Audit Metadata