The Agent Skills Directory

[COMMAND_EXECUTION]: Uses the Bash tool to execute a find and grep pipeline to identify existing markdown documentation files within the repository for migration purposes.
[DATA_EXFILTRATION]: Scans the project directory for configuration files (like package.json or pyproject.toml) and documentation to populate project metadata. It also specifies a check for secrets within the .claude/ directory to advise on .gitignore usage, though no external exfiltration logic is present.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted content from existing markdown documentation files to categorize and move them.
Ingestion points: Reads content from existing markdown documentation and project-specific configuration files during the migration and manifest creation steps.
Boundary markers: No delimiters or instructions to ignore embedded commands are present when processing the external file content.
Capability inventory: Includes file read/write operations, directory creation, and execution of Bash commands.
Sanitization: There is no evidence of sanitization or validation of the content found within the markdown files before it is analyzed for document type classification.

blueprint-init