The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script (search_bili_user.py) to perform the account search. The script is called with specific arguments for keywords and sorting, following standard automation practices.
[DATA_EXFILTRATION]: The skill communicates with https://redfox.hk/story/api/bili/userSearch to retrieve search results. This is a vendor-owned resource belonging to the skill author (redfox-data) and is necessary for the skill's primary functionality.
[PROMPT_INJECTION]: The skill fetches and processes Bilibili user biographies and nicknames to perform AI-based classification. This constitutes a surface for indirect prompt injection, as malicious instructions embedded in a Bilibili profile could attempt to influence the agent's categorization logic. However, this is a standard risk for skills processing external web content.
Ingestion points: Account nicknames and descriptions are processed in references/core_workflow.md after being fetched by scripts/search_bili_user.py.
Boundary markers: No specific delimiters or warnings to ignore embedded instructions are present in the processing step.
Capability inventory: The skill is capable of executing its own local search script but does not have broader system-level write access or administrative privileges.
Sanitization: No explicit sanitization of the external biography text is performed before it is passed to the AI for classification.

bilibili-keywords-accounts