linkfox-kalodata-tiktok-product

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns such as prompt injection, obfuscation, or unauthorized data exfiltration were detected. The skill's behavior aligns with its stated purpose of TikTok product research.\n- [COMMAND_EXECUTION]: The skill utilizes local Python scripts (scripts/kalodata_product_search.py, scripts/kalodata_product_detail.py, and scripts/onboarding.py) to process user requests and interface with vendor APIs. These scripts are invoked with JSON-formatted arguments and support local caching.\n- [EXTERNAL_DOWNLOADS]: The scripts perform network requests to recognized vendor domains (tool-gateway.linkfox.com, api.linkfox.com, agent-api.linkfox.com, and skill-api.linkfox.com) for data retrieval and user authentication. The onboarding.py script also provides instructions for installing common Python packages (requests, qrcode, pillow) if they are missing from the environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external API sources.\n
  • Ingestion points: Product ranking and detail data are ingested through the Kalodata API endpoints described in references/api.md.\n
  • Boundary markers: Data is handled as structured JSON; no explicit prompt delimiters are defined for the final output stage.\n
  • Capability inventory: The skill has the capability to write JSON response data and session metadata to the local linkfox/ workspace directory and perform network calls to vendor-controlled domains.\n
  • Sanitization: The scripts parse API responses as JSON, and the instructions recommend using tools like jq to extract specific fields rather than injecting entire payloads into the prompt context.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 12:21 PM
Security Audit — agent-trust-hub — linkfox-kalodata-tiktok-product