What problem does it solve?
Repositories often lose or scatter important conventions, design decisions, and agent behaviors across commits, PRs, and ephemeral AI sessions, making it hard to reproduce or follow standardized practices. This Skill captures those guidelines in a central, versioned location so future contributors and AI agents can discover and apply them consistently.
Core Features & Use Cases
- Intent distillation: Translate freeform user input into concise, technical guidelines suitable for documentation.
- Target resolution & safe writing: Determine the appropriate file (rules/, docs/conventions/, AGENTS.md, or .agents/.claude layers), detect conflicts, and write updates following repository resolution rules.
- Conflict analysis & synchronization: Halt and report contradictions for user clarification, update local skill files when applicable, and discover affected skills including symlink resolution for coordinated updates.
- Use Case: Record a new CI requirement or architectural decision so subsequent AI sessions and contributors reference the same, authoritative rule.
Quick Start
Use the remember skill to persist the guideline always use the --frozen-lockfile flag in CI.