The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data such as logs, exceptions, and stack traces, which serves as a potential injection surface.
Ingestion points: The skill instructions in SKILL.md specify the analysis of logs, exceptions, and telemetry data.
Boundary markers: There are no explicit instructions or delimiters defined to isolate the data being analyzed from the agent's operational instructions.
Capability inventory: The skill operates in a read-only sandbox as per the metadata and does not call for any dangerous tools, file-system modifications, or network operations.
Sanitization: No sanitization or input validation steps are described for the analyzed content.
[SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected. The skill instructions focus on evidence-based analysis and debugging best practices.

error-detective