The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs shell command execution to manage the documentation lifecycle.
Executes scripts/validate-frontmatter.sh to validate YAML headers against a schema.
Uses grep to search existing documentation for similar issues.
Uses mkdir -p, echo, and cat to create and update files in the docs/solutions/ directory.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection due to the way it processes and stores conversation content.
Ingestion points: Extracts module names, error messages, and investigation details from the conversation history in references/documentation-process.md (Step 2).
Boundary markers: Content is interpolated into documentation templates without explicit boundary markers or instructions to disregard potential commands in the source data.
Capability inventory: The workflow involves file-write capabilities (echo, cat) and local script execution (bash).
Sanitization: Filenames are sanitized for length and special characters; YAML metadata is validated using a regex-based bash script.

compound-docs