What problem does it solve? AI agents working in this repository often waste time on redundant builds, run destructive git commands, skip linting, or lose hard-won knowledge between sessions. This Skill establishes mandatory meta-instructions that govern every task, preventing costly mistakes and preserving learnings across sessions. ## Core Features & Use Cases - Auto-Knowledge Generation: Requires agents to check and append learnings to .agent/AI_GENERATED_KNOWLEDGE.md before and after complex tasks, building institutional memory. - Safe Verification Workflow: Mandates yarn ai:lint and yarn ai:test for changed-file-scoped linting and testing, while strictly prohibiting full builds and broad test suites. - Destructive Command Protection: Blocks git reset --hard, git clean -fd, rm -rf, and similar commands unless the user explicitly approves them. - PR Template Compliance: Enforces the .github/pull_request_template.md structure and preserves the PR-Agent AI summary block when creating pull requests. - Use Case: An agent fixing a bug in an Nx workspace automatically lints only changed files, runs colocated spec tests, records the root cause in the knowledge file, and drafts a compliant PR without touching uncommitted user work. ## Quick Start Apply the fundamental-principles rules to every task in this workspace, including scoped linting, knowledge logging, and PR template compliance.
