The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a markdown code block demonstrating how to use curl and cmp to detect 'soft-404' false positives. This is an educational example of how a human researcher would manually verify if a server returns the same response for a junk path as it does for a sensitive file. The commands are benign and use /tmp and $RANDOM safely.
[PROMPT_INJECTION]: No prompt injection or safety bypass patterns were detected. The instructions are focused on adherence to scope and preventing accidental attacks on third parties.
[DATA_EXFILTRATION]: No network exfiltration or sensitive data access patterns were found. The skill describes methodologies for analyzing existing datasets rather than creating new network connections or accessing local secrets.
[EXTERNAL_DOWNLOADS]: No external packages or remote scripts are referenced or downloaded.
[NO_CODE]: The skill is primarily documentation and methodology-based instructions without any integrated scripts or executable automation that would pose a runtime risk.

recon-scope-triage