The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a bash command in SKILL.md (ls -la <chapter-path>/*.md | grep -v summary | grep -v README | grep -v quiz) to locate lesson files for processing.
[INDIRECT_PROMPT_INJECTION]: The skill evaluates untrusted data from external markdown files as described in SKILL.md.
Ingestion points: The agent reads all .md files in a user-provided directory and references lessons from the local file system.
Boundary markers: No specific delimiters or "ignore instructions" markers are used when reading the contents of the files.
Capability inventory: The skill identifies files via ls and reads their content to perform linguistic and structural analysis.
Sanitization: No explicit sanitization or validation of the markdown content is performed before analysis.

chapter-evaluator