The Agent Skills Directory

[PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by ingesting untrusted codebase content and research data to generate specifications for subagents.
Ingestion points: Reads code, documentation, and research results from the repository and external searches during the self-discovery phase in SKILL.md.
Boundary markers: The instructions do not specify explicit delimiters or warnings to ignore embedded instructions when processing external content.
Capability inventory: The skill orchestrates subagents with build capabilities and executes system commands for testing and state management.
Sanitization: No explicit sanitization or filtering of external content is mentioned before interpolation into prompts.
[COMMAND_EXECUTION]: The skill uses shell commands to manage project state and verify work quality.
Evidence: The orchestrator executes typecheck, tests, git mv, and git commit to manage waves and verify implementation results.
[EXTERNAL_DOWNLOADS]: The skill provides installation instructions via the skills CLI from a vendor-owned repository.
Evidence: Instructions in README.md use npx -y skills to add the skill pack from the author's GitHub repository (yigitkonur/skills-by-yigitkonur). This is documented as a standard vendor resource.

run-aligned-delivery