with one click
archguard
archguard contains 6 collected skills from yaleh, with repository-level occupation coverage and site-owned skill detail pages.
Skills in this repository
Assemble Cognitive Context Bundles (CCBs) in parallel for all files in a planned edit set. Call before starting a multi-file editing session to pre-load cognitive load scores and edit precautions.
Run the 5-step Cognitive Analysis Loop to classify files as Pattern A/B/C and emit a per-package cognitive load heatmap. Pass a package path or file glob as the argument.
Converts a feature discussion into a single GitHub Issue for the L0 agent queue using a 5-phase iterative pipeline: (1) generate proposal draft, (2) add implementation details, (3) iteratively review proposal until APPROVED (max 3 attempts), (4) iteratively review goal-phase alignment until APPROVED (max 3 attempts), (5) create GitHub Issue without execution labels for human confirmation. No working-tree switching or remote-push operations.
Discover repository-specific project semantics and write them as `ProjectSemanticsInput` into `.archguard/project-semantics.json` or `archguard.config.json.projectSemantics`. Use when the task is to identify test discovery conventions, assertion wrapper patterns, Mermaid package-layer grouping, or other evidence-backed semantics that ArchGuard should consume.
Execute the full feature development lifecycle for a new feature or substantial change: generate a proposal from discussion or requirements, review and refine the proposal, generate and review an implementation plan, then implement the plan with TDD and validate the result. Use when given a feature discussion, requirements, or existing proposal/plan documents that should be turned into reviewed design docs and working code.
RLM (Refactoring Lifecycle Management): Systematic methodology for managing the complete lifecycle of software refactoring initiatives through 6 phases (PROPOSAL → PLANNING → EXECUTION → VALIDATION → INTEGRATION → MONITORING) with bi-directional traceability, quality gates, and metrics-driven measurement. Use when managing multiple refactoring iterations, improving plan-execution alignment, establishing refactoring dashboard and traceability matrix, or coordinating team refactoring efforts.