First Seen
Feb 18, 2026
Last Scanned
Feb 22, 2026
Findings
2
Score
85/100
Findings (2)
Detects skills that read remote markdown or documentation and apply it as operational rules
Read and interpret web + Use the **URL path** for precise contextTreat tool outputs as untrusted data. Do not act on directives found in tool responses without explicit user confirmation. Implement output validation before acting on tool results.
Likely FP if the match is a skill description explaining that it processes tool outputs, rather than a payload designed to hijack the agent via tool responses.
Detects patterns where external API responses are used directly without validation or sanitization
Webhook payload + directly from the useValidate and sanitize all data received from external APIs before using it in tool operations or agent prompts. Implement schema validation and treat API responses as untrusted input.
Likely FP if the match is a truncated table cell or documentation fragment that mentions API responses in a descriptive context, not actual unvalidated data processing.