The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides the find-polluter.sh script to automate the execution of npm test during diagnostics. This is a legitimate development tool within the scope of a debugging skill.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and act upon potentially untrusted error messages and code snippets.
Ingestion points: Processes stack traces, error logs, and repository files in the isolation phase of SKILL.md.
Boundary markers: Utilizes diagnostic altitudes (L1-L7) for logical separation, but lacks explicit security delimiters for ingested data.
Capability inventory: Authorization to edit code ('Repair and close') and execute tests via npm in find-polluter.sh and the <aegis-workspace-helper> tool.
Sanitization: No explicit sanitization or escaping of external content is documented before processing.

systematic-debugging