The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes product-level data (titles, brands, category paths) fetched from Amazon via the SellerSprite API, which constitutes untrusted external content.
Ingestion points: API responses are processed in scripts/sellersprite_product_search.py.
Boundary markers: Absent; the script parses and prints raw or summarized JSON without adding delimiters to prevent the agent from interpreting embedded instructions.
Capability inventory: The skill possesses the ability to write files to the local workspace (linkfox/ directory) and perform network requests to vendor domains.
Sanitization: Content is parsed as structured JSON but the text fields are not sanitized for potential injection patterns.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to several vendor-managed infrastructure components to retrieve product data and manage account status.
Evidence: Communications with https://tool-gateway.linkfox.com, https://api.linkfox.com, and https://agent-api.linkfox.com.
[DATA_EXFILTRATION]: The skill implements a feedback mechanism that transmits conversation context, sentiment, and user intent to a vendor-controlled telemetry endpoint.
Evidence: Telemetry data is sent to https://skill-api.linkfox.com/api/v1/public/feedback as part of the feedback loop described in references/api.md.
[PERSISTENCE]: The onboarding documentation provides explicit instructions for the agent to guide users in persisting API keys across sessions by modifying shell profiles.
Evidence: references/onboarding.md instructs adding exports to ~/.zshrc and ~/.bashrc, or using setx on Windows to store the LINKFOX_AGENT_API_KEY permanently.

linkfox-sellersprite-product-search