housing-scout

Pass

Audited by Gen Agent Trust Hub on Oct 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches real estate listing data from Redfin through the r.jina.ai proxy and environmental noise data from the Overpass API (overpass-api.de). Both are well-known technology services used for public data retrieval.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted external HTML and CSV data from real estate providers. This content is transformed into natural language summaries, creating a surface where malicious instructions hidden in listing descriptions could influence the agent's logic.
  • Ingestion points: scripts/housing_scout/providers/redfin.mjs (fetches via Jina proxy).
  • Boundary markers: Results are structured using markdown formatting, bullet points, and labeled headers (e.g., 'Price:', 'Summary:') in scripts/housing_scout/housing_scout.mjs.
  • Capability inventory: The skill utilizes the exec tool to run local scripts, performs network requests via fetch, and writes to local JSON files in the data/ directory.
  • Sanitization: Content is processed using regular expressions to extract specific numerical and text fields, which limits the direct passage of raw external strings but does not fully neutralize embedded instructions.
  • [COMMAND_EXECUTION]: Relies on the exec tool to run its core logic via a bundled Node.js script (housing_scout.mjs).
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 12, 2026, 05:35 PM