The Agent Skills Directory

[DATA_EXFILTRATION]: The skill performs network operations to retrieve landing page content for analysis. Evidence: scripts/main_tool.py uses the requests library to fetch data from URLs provided by the user.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted text from external websites, which is an inherent surface for indirect prompt injection in analysis tools. 1. Ingestion points: scripts/main_tool.py extracts text and metadata from remote URLs. 2. Boundary markers: The agent instructions do not explicitly define delimiters to separate untrusted content. 3. Capability inventory: The skill is limited to reading data and providing analysis; it lacks capabilities for writing files or executing arbitrary system commands. 4. Sanitization: The skill uses BeautifulSoup to extract plain text but does not implement specific sanitization for malicious prompt patterns.

web3-ux-feedback