The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted data found within project source files and design documentation.- Ingestion points: Analyzes UI source code (extensions .tsx, .ts, .css) and design specifications (e.g., DESIGN.md) as defined in SKILL.md.- Boundary markers: The skill does not define specific delimiters for processed data but uses structured checklists to guide the analysis.- Capability inventory: Utilizes shell access for git operations and the 'herdr' CLI for agent lifecycle management.- Sanitization: No explicit sanitization of file content is performed; however, the skill mandates a strict 'Read only' policy for all agents, specifically forbidding the execution of tests, builds, or database interactions.- [COMMAND_EXECUTION]: The skill uses local shell commands to facilitate development workflows.- Evidence: Executes 'git diff' to identify changed files and uses the 'herdr' toolset for agent orchestration and status monitoring (e.g., 'wait-all-idle.sh'). these commands are used for their intended purpose in a development environment and do not involve remote execution or downloading external scripts.

design-review