E2E verification with intent-aware scenario design and multi-probe evidence chains. Supports TWO modes: (A) live server testing with config manipulation, (B) integration test authoring for library/core modules. Auto-detects mode from project structure. Use when: (1) user says "verify", "e2e test", "production verify", "validate module", "生产验证", "验证模块", (2) after completing a major feature that adds endpoints, tools, or code paths, (3) user runs /e2e-verify [target]. Analyzes git diff + user instructions to design positive/negative/boundary scenarios. Flags: --debug (verbose output), --skip-build (reuse last binary).
Audit Aleph's knowledge graph (`.understand-anything/knowledge-graph.json`) for code-health signals — orphan nodes, dead-code candidates, R1 brain-limb separation violations, R10 thin-harness budget overrun. Use when the user types `/graph-review`, asks to "audit the graph", "check redlines", "find orphans", "find dead code", "verify thin harness", or wants a code-health pulse before a release / refactor / merge. Requires the graph file to exist — run `/understand` first if missing. Wraps `scripts/graph-audit.mjs` (no LLM cost, pure deterministic JSON walk).
Logic-level code review checklist (Aleph local skill). STUB — fill in concrete trigger phrases, scope, and step-by-step procedure before relying on this skill. See docs/rust-logic-audit-prompts.md for seed material.
Deep logic bug review for Rust codebases. Portable engine for logic chain analysis including: (1) Topology Reconnaissance, (2) Context Alignment, (3) Semantic Invariant checking, (4) Connectivity & Wiring audit, (5) Control flow simulation, (6) Red-teaming, (7) Five-layer automated verification (L1-L5). Trigger: /rust-logic-audit [module|commit] [--strict]
Basic echo command (OpenClaw compatible)
GitHub PR operations (Aleph enhanced with Docker)