The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands via curl to retrieve HTTP headers and page content from user-supplied URLs. It also uses grep for pattern matching within the fetched content.
[EXTERNAL_DOWNLOADS]: The skill fetches content from external URLs and specifically targets JavaScript libraries from well-known services and CDNs, including jsdelivr.com, cdnjs.com, unpkg.com, and googleapis.com for analysis.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted HTML and JavaScript content from external websites, creating a surface for indirect prompt injection. Malicious instructions embedded in a target website could attempt to influence the agent's subsequent analysis or report generation.
Ingestion points: The skill uses curl to download page bodies and external JavaScript files into the agent's processing environment.
Boundary markers: Absent; there are no explicit instructions for the agent to treat the fetched content as untrusted data or to ignore embedded instructions.
Capability inventory: The agent has access to Bash, WebFetch, Read, Grep, and Glob tools.
Sanitization: Absent; the skill performs direct parsing and interpretation of raw external content without prior sanitization.

attack-surface-xss