The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains an architectural surface for indirect prompt injection because it processes untrusted system logs and stack traces while maintaining access to the shell. \n
Ingestion points: The skill is designed to read and analyze logs, stack traces, and production error data from the filesystem (SKILL.md). \n
Boundary markers: The investigation protocol does not include instructions for using delimiters or boundary markers to isolate processed data from the agent's command context. \n
Capability inventory: The skill utilizes Bash, Grep, Read, and Glob to fulfill its purpose, which involves system-level file access and shell command execution (SKILL.md frontmatter). \n
Sanitization: No explicit sanitization or filtering logic is specified for the external log data before it is interpreted by the agent.

error-detective