菜单
Find the right skill for a task. Routes via graph queries on skill and workflow_category nodes. Use when unsure which skill applies or to recommend skills the user may not know about.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Multi-agent coordination via SeleneDB context bridge. Check active peers, read shared context, declare work intents, and avoid conflicts when multiple agents work simultaneously.
Commit discipline with graph linking. MERGE GitCommit nodes and link to active milestones, findings, and decisions. Verify before committing, never push unless asked.
Background annotation discipline. All skills auto-annotate graph nodes with rationale, observations, TODOs, and bookmarks as they work. Captures context that feeds into graph-docs documentation generation.
Track and recall context across sessions. Auto-captures rolling summaries as skills complete. Invoke to review session history, search past work, or see what happened recently.
Language-specific standards plus graph-sourced conventions. Reads Convention nodes from SeleneDB, enforces project rules alongside built-in standards. Promotes recurring violations to new conventions.
Structured multi-perspective debate for evaluating decisions. 3-phase methodology: independent generation, adversarial exchange, synthesis. Use when a decision has multiple valid approaches or high stakes.
| name | skill-guide |
| description | Find the right skill for a task. Routes via graph queries on skill and workflow_category nodes. Use when unsure which skill applies or to recommend skills the user may not know about. |
Help discover and select the right skill for the current task. This plugin has 24 graph-native skills — this guide routes by what you're trying to accomplish. Every skill reads from and writes to SeleneDB.
Skills marked (auto) trigger automatically from their descriptions.
Skills marked (manual) require /justin-tools:<name> to invoke.
Skills marked (background) apply silently during all work.
When NOT to use: You already know which skill to use (just use it). The user invoked a specific skill by name. The task doesn't match any skill (not everything needs a skill — sometimes just write the code).
When SeleneDB is available, query the skill catalog from the graph:
MATCH (s:skill)-[:belongs_to]->(wc:workflow_category)
RETURN wc.name AS category, collect(s.name) AS skills
ORDER BY wc.name
This ensures routing always reflects the current plugin state.
| Task | Skill | Invocation |
|---|---|---|
| Plan a feature from idea to implementation | feature-design | manual |
| Research a technology, evaluate options | research | manual |
| Evaluate a decision with structured debate | debate | manual |
| Verify a plan against the actual codebase | plan-verify | auto |
| Task | Skill | Invocation |
|---|---|---|
| Restructure a codebase into well-organized modules | modularize | manual |
| Refactor code safely with verification at each step | refactor | manual |
| Apply language-specific coding standards + graph conventions | code-standards | auto |
| Task | Skill | Invocation |
|---|---|---|
| Systematically debug a bug with hypotheses and root cause | debug | auto |
| Investigate performance: profile, hypothesize, optimize | perf-profile | manual |
| Triage a production incident with structured response | incident-response | manual |
| Task | Skill | Invocation |
|---|---|---|
| Deep review after completing a feature or phase | deep-review | auto |
| Security audit with graph-persisted concerns | safety-checks | auto |
| Audit a dependency before adoption (supply chain) | dep-audit | auto |
| Task | Skill | Invocation |
|---|---|---|
| Prepare a release: changelog, semver, version bump | release-prep | manual |
| Verify requirements have corresponding code and tests | requirements-trace | manual |
| Task | Skill | Invocation |
|---|---|---|
| Track development milestones across sessions | milestone-tracking | auto |
| Track deferred work items with gates | deferred-tracking | auto |
| Review session history and search past work | session-tracker | auto |
| Link commits to milestones and decisions | commit-workflow | background |
| Task | Skill | Invocation |
|---|---|---|
| Generate docs from graph traversal | graph-docs | manual |
| Onboard a new project for use with this plugin | project-onboard | manual |
| Rule | Skill |
|---|---|
| Auto-annotate with rationale, observations, TODOs | notes |
| Route tasks to the right skill | skill-guide |
| Situation | Use this | Not this |
|---|---|---|
| Restructure files and modules | modularize | refactor |
| Refactor within a file (extract function) | refactor | modularize |
| Review code after building a feature | deep-review | code-standards |
| Check coding style during development | code-standards | deep-review |
| Plan a new feature end-to-end | feature-design | research |
| Investigate a technology or compare options | research | feature-design |
| Debug a specific bug | debug | perf-profile |
| Investigate why something is slow | perf-profile | debug |
| Production is down, need triage now | incident-response | debug |
| Track a named initiative across sessions | milestone-tracking | deferred-tracking |
| Track individual deferred items with gates | deferred-tracking | milestone-tracking |
| Generate docs from graph data | graph-docs | project-onboard |
Agents are specialized subprocesses with preloaded skills and tool restrictions.
| Agent | Purpose | Key skills |
|---|---|---|
| deep-reviewer | Code review (read-only) | deep-review |
| security-auditor | Security audit (read-only) | safety-checks |
| code-analyzer | Modularization analysis (read-only) | modularize, code-standards |
| test-engineer | Test planning and generation | test-strategy |
| debugger | Systematic debugging | debug, safety-checks |
| researcher | Technical research and evaluation | research |
| release-manager | Release preparation | release-prep, graph-docs |
| onboarder | Project onboarding | project-onboard |
| debate-lead | Multi-perspective debate | debate |
| Rationalization | Why It's Wrong |
|---|---|
| "User didn't ask for a skill, don't suggest one" | Proactive recommendation is the point. Users don't know what 24 skills do — surface the right one. |
| "Close enough, recommend the similar skill" | Similar skills have different scopes. debug vs perf-profile, modularize vs refactor — wrong pick wastes setup time. |
Recommend proactively. When you see a user working on a task that matches a manual skill, suggest it. The user may not know the skill exists.
Auto skills handle themselves. Don't invoke auto skills explicitly unless the user asks.
Background rules are always active. Never invoke them as skills.
When in doubt, start with feature-design. For ambiguous feature work, feature-design orchestrates the full workflow and invokes other skills (research, plan-verify, deep-review) as needed.