bing-image-search-mcp

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFECREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The scripts 'scripts/search_images.py' and 'scripts/search_images_batch.py' contain a hardcoded demo access key ('BETA_TEST_KEY_MARCH_2026'). While intended for testing, hardcoding credentials is a suboptimal practice.
  • [EXTERNAL_DOWNLOADS]: The skill connects to the vendor's API endpoint at 'deepnlp.org' via the 'ai_agent_marketplace' library.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection.
  • Ingestion points: The 'query' and 'query_list' parameters in 'scripts/search_images.py' and 'scripts/search_images_batch.py' ingest user-controlled data.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the scripts or metadata.
  • Capability inventory: The skill communicates with an external API service (deepnlp.org) and returns the retrieved data to the user.
  • Sanitization: No input validation, escaping, or filtering is applied to the tool arguments before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 08:45 AM
Security Audit — agent-trust-hub — bing-image-search-mcp