ngx-ai-devtools-angular-llm-debugger

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of the ngx-ai-devtools package from the NPM registry to function.
  • [DATA_EXFILTRATION]: As its primary purpose, the tool intercepts outgoing network requests to AI providers (OpenAI, Anthropic, Gemini, etc.). It captures sensitive headers (e.g., Authorization tokens), prompts, and responses. This data is stored in the application's memory and can optionally be saved to the browser's localStorage for persistence across sessions.
  • [COMMAND_EXECUTION]: Provides a replay() capability that allows developers to programmatically re-issue captured network requests to LLM providers.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes and renders untrusted data from external API responses.
  • Ingestion points: Intercepted HTTP response bodies and streaming event deltas from LLM APIs (SKILL.md).
  • Boundary markers: None identified; the tool displays raw intercepted content in a UI panel.
  • Capability inventory: Includes the ability to re-trigger network requests via a replay() function.
  • Sanitization: Relies on the host Angular application's standard rendering and security practices for the devtools UI components.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 05:09 PM
Security Audit — agent-trust-hub — ngx-ai-devtools-angular-llm-debugger