The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's core workflow involves auditing and refactoring existing source code, which serves as an ingestion point for potentially untrusted data.
Ingestion points: The skill uses read_file and search_code to scan component prop surfaces and implementation logic as described in SKILL.md.
Boundary markers: There are no explicit instructions for the agent to use delimiters or 'ignore' prefixes when processing content from these files.
Capability inventory: The skill has access to sensitive capabilities including file modification (write_file, edit_file) and shell command execution (execute_command, run_tests, run_build).
Sanitization: No specific sanitization or validation requirements are defined for the code content before it is processed by the model.

component-composition