linkfox-amazon-policy-feed

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill connects to tool-gateway.linkfox.com, api.linkfox.com, and agent-api.linkfox.com to fetch policy updates and manage user authentication. These endpoints are owned and operated by the skill's author (linkfox-ai).
  • [COMMAND_EXECUTION]: The tool invokes local Python scripts (amazon_policy_feed.py, amazon_policy_feed_detail.py, and onboarding.py) to handle API interactions and user onboarding processes.
  • [DATA_EXFILTRATION]: API results are stored in a linkfox/ directory within the project workspace. This is a standard persistence pattern for this vendor, allowing the agent to reference historical policy data without re-triggering API calls.
  • [INDIRECT_PROMPT_INJECTION]: As the skill retrieves and processes external content from Amazon's policy feed, it is subject to indirect prompt injection risks if the source content contains malicious instructions.
  • Ingestion points: External data is ingested through the amazon/policyFeed and amazon/policyFeedDetail scripts described in SKILL.md.
  • Boundary markers: There are no explicit boundary delimiters or warnings to ignore instructions within the ingested policy text.
  • Capability inventory: The skill possesses network access to the vendor's gateway and write access to the project's local file system.
  • Sanitization: The scripts utilize standard JSON parsing for API responses; policy article details are rendered for the agent as Markdown.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 03:42 PM
Security Audit — agent-trust-hub — linkfox-amazon-policy-feed