The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon instructions found in external specification files (spec README) and "skills named by the spec".
Ingestion points: The workflow starts by reading repository READMEs, spec READMEs, and loading referenced skills (SKILL.md, Step 1).
Boundary markers: The instructions do not define explicit delimiters or "ignore" instructions for content sourced from these external files.
Capability inventory: The skill has the ability to modify the codebase, execute git commands, spawn concurrent subagents, and perform browser-based verification (SKILL.md, Rules and Workflow sections).
Sanitization: The workflow incorporates iterative review processes, including review, refactor-clean, and audit-choices skills, to validate the code and decisions made during implementation.

implement-spec