The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a utility script find-polluter.sh which uses the find command and executes npm test on identified files to isolate tests causing environmental side effects.
[COMMAND_EXECUTION]: The SKILL.md file provides instructional examples for diagnostic instrumentation, suggesting the use of shell commands such as env, security list-keychains, and codesign to verify environment state during debugging.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection (Category 8) as its core process involves the agent ingesting and analyzing potentially untrusted error messages and stack traces.
Ingestion points: Technical error messages, stack traces, and component logs ingested in Phase 1 (Root Cause Investigation).
Boundary markers: Absent; the skill does not define specific delimiters for separating error data from instructions.
Capability inventory: File system access, shell command execution via the provided bash script, and local environment inspection.
Sanitization: None; the skill assumes the technical data (errors/logs) is safe for analysis.

systematic-debugging