xoai/sage - 第 2 页

sage

Start here. Sage reads project state, routes via keywords, classifies intent, and guides you to the right workflow.

原文语言：英语

sage-navigator

Activates when the user starts any substantial task: building, creating, redesigning, analyzing, researching, planning, fixing, improving, evaluating, writing, auditing — code, products, content, or strategy. Also activates when the user asks what to do next,…

原文语言：英语

autoresearch

Autonomous iteration toward a measurable outcome. Use when the user wants to optimize a numeric metric through repeated modify-verify cycles — reduce bundle size, increase test coverage, improve query time, lower readability score. Not for exploratory…

原文语言：英语

onboard

First-run project setup that detects tech stack, selects quality packs, and generates .sage/ directory with CLAUDE.md. For new projects, guides technology selection. Use when no .sage/ directory exists, when the user says "set up sage", "initialize", "get…

原文语言：英语

baas

"Corrects the most common BaaS architecture mistakes agents make — security rules as authorization, data modeling for queries, minimizing serverless functions, real-time by default, and client-direct access patterns. Applies to Firebase, Supabase, and similar…

原文语言：英语

flutter

Flutter patterns — widget architecture, state management, Impeller renderer, platform-adaptive design

原文语言：英语

nextjs

Next.js 14/15 App Router patterns — server components, data fetching, caching, server actions

原文语言：英语

react-native

React Native patterns — New Architecture (Fabric/TurboModules), Expo, navigation, performance

原文语言：英语

learn

Knowledge entries in sage-memory, Docs in .sage/docs/

原文语言：英语

reflect

Cycle review, Learnings with prevention rules, Next-cycle seeds

原文语言：英语

review

Independent artifact review via sub-agent delegation. Evaluates completeness, consistency, and quality with severity classification.

原文语言：英语

sage-memory

Integrates sage-memory into Sage workflows. Teaches the agent when to remember (store findings during work), when to recall (search memory at session start and task start), and how to learn (structured knowledge capture via sage learn). Use when the user…

原文语言：英语

git-discipline

Branch-per-initiative git discipline for all delivery workflows. Defines branch naming by workflow, the propose-confirm creation protocol, dirty-tree and detached-HEAD handling, the always user-gated merge protocol, worktree support for parallel sessions, and…

原文语言：英语

coding-principles

Use during implementation — loaded by build-loop before each task — whenever code is being written and should hold to universal quality standards, regardless of language or stack. Active while writing, not a post-hoc review checklist.

原文语言：英语

autonomous

When --autonomous is active, skip user-facing elicitation rounds. The agent makes elicitation decisions from memory, codebase patterns, constitution principles, and prior cycle context. Every decision cites its source. Substantive decisions with no signal…

原文语言：英语

testing-skills

Use when authoring or editing a skill that enforces a rule an agent could rationalize past — a discipline skill — or when a shipped discipline skill is being skipped, ignored, or only partially followed. Also use when reviewing a discipline skill before…

原文语言：英语

skill-no-marker

Fixture discipline skill that is MISSING its compliance_marker. The harness must exit 2 (setup error) when asked to run this.

原文语言：英语

skill-no-tests

Fixture discipline skill with a marker but NO sibling TESTS.md. The harness must exit 2 (setup error) because there is no scenario to dispatch.

原文语言：英语

skill-ok

Fixture discipline skill used by run-skill-test.test.sh. Declares a real compliance marker so the harness has something to grep for.

原文语言：英语

deep-elicit

Comprehensive requirements elicitation through Socratic conversation, producing a product brief with problem, users, scope, and success criteria in ~10 minutes. Use when starting a new product, designing a system from scratch, planning a major redesign, or…

原文语言：英语

quick-elicit

Guided specification through 4 focused question rounds in ~3 minutes. Round 0 challenges problem framing before solutioning. Produces framing, intent, boundaries, and acceptance criteria. Use when the user wants to add a feature, build something new,…

原文语言：英语

browser-check

Quick browser smoke test via Lightpanda MCP. Checks if the primary affected route renders, has no JS errors, and contains expected elements. Advisory only — never blocks. Invisible when Lightpanda is not available or no frontend files in diff.

原文语言：英语

design-check

Quick design quality scan of changed frontend files. Checks for hardcoded colors (when design system exists), missing interactive states, and AI slop indicators. Code-only — does not use Lightpanda. Advisory only — never blocks. Invisible when no frontend…

原文语言：英语

context-loader

Defines how Sage content is loaded into the agent's context window. This capability guides platform generators — it specifies what to inline, what to reference, and what to skip. Not a runtime skill for agents; a build-time strategy for generators.

原文语言：英语

tool-use

Guides tool selection across three layers: local scripts (zero context cost), MCP proxy (minimal context), and subagent dispatch (isolated context). Ensures the agent uses the cheapest effective layer and never pollutes the main context window with raw tool…

原文语言：英语

implement

Implements a single task from the plan using TDD discipline. Reads the task spec, writes tests first, writes minimal passing code, refactors, and commits. Use when executing a planned task, writing code for a feature, or when the user says "implement this",…

原文语言：英语

sage-help

Provides specific, actionable guidance on what to do next based on current project state. Reads progress files and context to give one clear next step. Use when the user says "help", "what do I do", "sage help", "what's next", "I'm stuck", or "status".

原文语言：英语

api

Corrects the 13 most common API design mistakes agents make — grounded in Geewax, Amundsen, Ousterhout, Kleppmann, and Gough/Bryant

原文语言：英语

mobile

Universal mobile development principles — offline-first, 60fps, touch, battery, platform patterns

原文语言：英语

react

React 18/19 patterns and anti-patterns — hooks discipline, component architecture, state management

原文语言：英语

web

Universal web development principles — accessibility (WCAG 2.2), performance (Core Web Vitals), security headers (OWASP), SEO, responsive design, error UX, and loading states. Framework-agnostic.

原文语言：英语

systematic-debug

Four-phase debugging framework that finds root causes before attempting fixes. Use when investigating errors, debugging failures, fixing bugs, analyzing test failures, diagnosing unexpected behavior, examining stack traces, or when the user says "it's…

原文语言：英语

verify-completion

Verifies implementations and fixes actually work by executing commands and observing real results. Never trusts claims. Use as the final quality check after implementation, after a bug fix, or when the user asks "is it done", "does it work", "verify this", or…

原文语言：英语

codebase-scan

Scans relevant codebase areas to understand existing patterns, conventions, dependencies, and architecture before making changes. Use at the start of any feature or fix, when entering an unfamiliar codebase area, or when the user says "scan the project",…

原文语言：英语

specify

Structures elicitation output into a formal specification defining WHAT to build and WHY. Use after quick-elicit or deep-elicit completes, or when the user provides requirements and says "write a spec", "define requirements", "create a PRD", or "specify this…

原文语言：英语

visual-review

Evaluates UI implementation by analyzing screenshots at multiple breakpoints against the spec, design conventions, and web standards. Checks layout, hierarchy, responsiveness, spacing, and accessibility. Use after implementing UI components, after a redesign,…

原文语言：英语