The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/generate_migration_script.py is designed to dynamically generate shell scripts (.sh) based on templates and content provided at runtime.
Evidence: The generate_script method in scripts/generate_migration_script.py assembles a bash script from a hardcoded template and an input template string.
[PRIVILEGE_ESCALATION]: The skill programmatically modifies file system permissions to make its generated output files executable.
Evidence: file_path.chmod(0o755) in scripts/generate_migration_script.py ensures that any generated shell script is assigned execute permissions, enabling it to be run as code.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process database schema definitions from external environments, which serves as a potential attack surface for indirect prompt injection.
Ingestion points: SQL schema files (schema_source.sql, schema_target.sql) and results from information_schema queries.
Boundary markers: No delimiters or "ignore instructions" warnings are specified in the instructions or scripts for handling extracted schema data.
Capability inventory: The skill has the capability to write files, modify file permissions (chmod), and execute database tools via Bash.
Sanitization: The generate_migration_script.py script performs no validation or escaping of the content before interpolating it into executable script templates.

comparing-database-schemas