The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from TikTok livestream leaderboards and details (e.g., livestream titles, creator usernames). This creates a potential surface for indirect prompt injection if malicious instructions are embedded within the fetched metadata that the agent later processes.
Ingestion points: Data is fetched and written to local JSON files by scripts/livestream_detail.py and scripts/livestream_rank.py.
Boundary markers: No specific delimiters or "ignore instructions" warnings are applied to the external data before it is read by the agent.
Capability inventory: The skill utilizes file system writing and network requests via the vendor's API proxy.
Sanitization: No explicit filtering or sanitization of the external JSON content is performed.
[COMMAND_EXECUTION]: The skill instructions direct the agent to execute provided Python scripts (scripts/livestream_detail.py and scripts/livestream_rank.py) to perform its data fetching and ranking operations.
[DATA_EXFILTRATION]: The Python scripts perform network operations to the vendor's API proxy endpoint (defined by the NEXSCOPE_PROXY_BASE environment variable) to retrieve TikTok analytics. These requests utilize a user-provided NEXSCOPE_API_KEY for authentication.

ecommerce-tiktok-livestream-analytics