Skills de Claude Code · página 152

Consult an agent-wiki for guidelines relevant to the task at hand. The wiki itself documents how to retrieve from it (AGENTS.md). Use this skill once you know what task or sub-task you're about to do — not at session start.

Read a normalized Claude Code trajectory JSON and extract reusable guidelines into wiki-twobatch/guidelines/. Use when mining saved trajectories for reusable lessons.

Ingest one or more agent trajectories (raw bob/claude traces or normalized JSON) into an agent-wiki end-to-end — convert, summarize, extract guidelines, synthesize skills, consolidate into clusters, and catalog. Use when you have a batch of traces to turn into a wiki in one pass.

Read a normalized Claude Code trajectory JSON and write an episodic summary page to wiki-twobatch/summaries/. Use when summarizing one or more saved trajectories into the agent wiki.

Read a normalized Claude Code trajectory JSON and produce a wiki-resident SKILL.md page that future agents can invoke. Use when a trajectory captured a non-trivial successful workflow worth promoting from a free-text guideline to an executable, callable artifact.

Discover task families across summaries and write per-family comparison pages with findings narrative. Updates wiki-twobatch/_config.yaml task definitions and writes tasks/<slug>__task.md.

Must be used near the end of any non-trivial turn that produced potentially reusable tools, guidance, errors, workarounds, or workflows, so those lessons are saved for future turns.

Analyze saved trajectories and recall audit events offline to record whether recalled guidelines influenced completed sessions.

Publish a private guideline to a configured write-scope repo.

Must be used at the start of any non-trivial task involving code changes, debugging, repo exploration, file inspection, or environment/tooling investigation to surface stored guidance before analysis or tool use.

Save the current conversation as a trajectory JSON file in OpenAI chat completion format for analysis and fine-tuning

Captures the current session's successful workflow and saves it as a reusable skill with SKILL.md and helper scripts

Add a shared guidelines repo (read-scope subscription or write-scope publish target) to the unified repos list.

Pull the latest guidelines from every configured repo (read- and write-scope).

Remove a repo from the unified repos list and delete its local clone.

Must be used near the end of any non-trivial turn that produced potentially reusable tools, guidance, errors, workarounds, or workflows, so those lessons are saved for future turns.

Analyze saved trajectories and recall audit events offline to record whether recalled guidelines influenced completed sessions.

Publish a private guideline to a configured write-scope repo.

Must be used at the start of any non-trivial task involving code changes, debugging, repo exploration, file inspection, or environment/tooling investigation to surface stored guidance before analysis or tool use.

Save the current conversation as a trajectory JSON file in OpenAI chat completion format for analysis and fine-tuning

Captures the current session's successful workflow and saves it as a reusable skill with SKILL.md and helper scripts

Add a shared guidelines repo (read-scope subscription or write-scope publish target) to the unified repos list.

Pull the latest guidelines from every configured repo (read- and write-scope).

Remove a repo from the unified repos list and delete its local clone.

Build, modify, debug, and deploy agents with Agentforce Agent Script. TRIGGER when: user creates, modifies, or asks about .agent files or aiAuthoringBundle metadata; changes agent behavior, responses, or conversation logic; designs agent actions, tools, subagents, or flow control; writes or reviews an Agent Spec; previews, debugs, deploys, publishes, or tests agents; uses Agent Script CLI commands (sf agent generate/preview/publish/test). DO NOT TRIGGER when: Apex development, Flow building, Prompt Template authoring, Experience Cloud configuration, or general Salesforce CLI tasks unrelated to Agent Script.

Analyze production Agentforce agent behavior using session traces and Data Cloud. TRIGGER when: user queries STDM session data or Data Cloud trace records; investigates production agent failures, regressions, or performance issues; asks about session traces, conversation logs, or agent metrics; wants to reproduce a reported production issue in preview; runs findSessions or trace analysis queries. DO NOT TRIGGER when: user creates, modifies, or debugs .agent files during development (use developing-agentforce); writes or runs test specs (use testing-agentforce); uses sf agent preview for local development iteration; deploys or publishes agents.

Run OWASP LLM Top 10 security assessments against live Agentforce agents. TRIGGER when: user asks for security testing, OWASP scan, red-teaming, penetration testing, security grade, vulnerability assessment, prompt injection test, data leakage test, excessive agency test, security posture check, or hardening recommendations. DO NOT TRIGGER when: user runs functional smoke tests or batch tests (use testing-agentforce); performs static safety review of .agent file content (use developing-agentforce Section 15); analyzes production session traces (use observing-agentforce); writes or modifies .agent files.

Write, run, and analyze structured test suites for Agentforce agents. TRIGGER when: user writes or modifies test spec YAML (AiEvaluationDefinition); runs sf agent test create, run, run-eval, or results commands; asks about test coverage strategy, metric selection, or custom evaluations; interprets test results or diagnoses test failures; asks about batch testing, regression suites, or CI/CD test integration. DO NOT TRIGGER when: user creates, modifies, previews, or debugs .agent files (use developing-agentforce); deploys or publishes agents; writes Agent Script code; uses sf agent preview for development iteration; analyzes production session traces (use observing-agentforce); requests OWASP, security, or red-team testing (use securing-agentforce).

>

>

>

>

>

>

Use when setting up a project environment — installing dependencies, verifying builds, detecting the tech stack. Covers Phase 0 of a new session.

Use when committing code, pushing branches, writing .gitignore entries, or generating PR metadata. Covers commit format, branch rules, and artifact exclusion.

Use when updating project documentation, saving learnings, or running retrospectives. Covers CLAUDE.md updates, memory persistence, and cross-session learning.

>

>

Unified skill hub for Solana development. Routes to external submodule skills (solana-foundation, sendai, solana-game, trailofbits, cloudflare, qedgen, colosseum, solana-new, ghostsecurity, defending-code) and local skills. Progressive disclosure — read only what you need.

Prepare a winning hackathon submission. Use when the user says "hackathon submission", "submit to hackathon", "demo script", "demo video", "which track should I enter", "Colosseum", "help me win the hackathon", or asks about hackathon grants and Superteam Earn.

Find and validate what to build in crypto. Use when the user asks "what should I build", "validate this idea", "is this worth building", "find me a startup idea", "crypto idea", or wants blunt feedback on a project concept before writing code.

Build a pitch deck for a crypto project. Use when the user says "pitch deck", "demo day", "investor presentation", "grant application slides", "accelerator application", "help me pitch", or needs slides for a hackathon final.

Skill hub for the solana-ai-kit Claude Code plugin. Routes the bundled go-to-market skills and the opt-in add-on catalog, and points to upstream marketplaces (and the install.sh full install) for protocol, security, and ecosystem depth. Progressive disclosure — read only what you need.

LLM-powered injection of project context into installed agent templates via `aspens customize agents`

>

Core conventions, tech stack, and project structure for aspens

Claude/Codex CLI execution layer — prompt loading, stream-json parsing, file output extraction, path sanitization, skill file writing, and skill rule generation

Top-level Commander wiring, welcome screen, missing-hook warning, CliError exit handling, and the public programmatic API surface

Multi-target output system — target abstraction, backend routing, content transforms for Codex CLI and future targets

Context health analysis — freshness, domain coverage, hub surfacing, drift detection, LLM-powered interpretation, and auto-repair for generated agent context

Incremental skill updater that maps git diffs to affected skills and optionally auto-syncs via a post-commit hook

Static import analysis that builds dependency graphs, domain clusters, hub files, git churn hotspots, and file priority rankings

Deterministic repo analysis — language/framework detection, structure mapping, domain discovery, health checks, and import graph integration

Token-saving session automation — statusline, prompt guard, precompact handoffs, session rotation, and handoff commands for Claude Code

LLM-powered generation pipeline for Claude Code skills and CLAUDE.md — doc-init command, prompt system, context building, and output parsing

Bundled agents, commands, hooks, and settings that users install via `aspens add`, `aspens doc init`, and `aspens save-tokens` into their .claude/ directories

为 AI Agent 友好的代码库搭建和改进 Harness 工程（包括 AGENTS.md、docs/、Lint 规则、Eval 系统、项目级 Prompt 工程）。触发场景：为 AI Agent 设置新项目/空项目，创建 AGENTS.md 或 CLAUDE.md，关于 Harness 工程的问题，让 Agent 在代码库上更高效地工作。当用户感到沮丧或抱怨 Agent 质量时也会触发（例如：'Agent 总是无视规范'、'它从不听从指令'、'为什么它总是做错 X'、'Agent 坏了'）— 因为 Agent 输出质量差几乎总是意味着 Harness 缺失，而不是模型问题。涵盖：Context 工程、架构约束、多 Agent 协作、评估、长运行任务 Harness 以及 Agent 质量问题诊断。

Set up and improve harness engineering (AGENTS.md, docs/, lint rules, eval systems, project-level prompt engineering) for AI-agent-friendly codebases. Triggers on: new/empty project setup for AI agents, AGENTS.md or CLAUDE.md creation, harness engineering questions, making agents work better on a codebase. ALSO triggers when users are frustrated or complaining about agent quality — e.g. 'the agent keeps ignoring conventions', 'it never follows instructions', 'why does it keep doing X', 'the agent is broken' — because poor agent output almost always signals harness gaps, not model problems. Covers: context engineering, architectural constraints, multi-agent coordination, evaluation, long-running agent harness, and diagnosis of agent quality issues.

Find context from past Claude Code (CLI) and Claude Cowork (desktop) sessions on this Mac. Use when the user wants to recall something they did before but can't find it , phrasings like "where did I work on X", "find that session where I…", "when did I last do Y", "pull up the conversation about Z", "that time I built/tried/discussed …". Searches by kind (code/cowork), time range, title, working directory, or free-text content across all transcripts.

Mine the user's Claude Code + Cowork session history into a structured task profile, what they do with AI, how often, how successfully where friction lives, then propose atomic skills that would reduce iteration. Use when the user asks to "analyse my Claude use", "build a task profile", "what tasks do I do with Claude", "where am I spending tokens", "what skills would help me", or mentions reviewing past sessions for patterns. Produces profile.csv (shareable), explorer.html (personal coaching view with AI-first principle comparison + token-spend chart), and skill-proposals.md.

Personal diagnosis of where your Claude Code + Cowork spend goes. Reads local transcripts, prints your conversation length distribution, marathon share, cache rebuild costs, and per-project diagnosis (good projects and problem projects) right in the terminal. Then offers a deeper dive that fans out parallel Haiku subagents over your most expensive (and most efficient) sessions and writes a tight Markdown report. Use when the user asks "why is my Claude spend so high", "where am I burning tokens", "diagnose my Claude habits", "audit my Claude usage", or asks for a personal token-cost diagnosis.

Analyze, re-engineer, or bootstrap projects to align with AI-first design principles. Use when asked to review, audit, improve, 'ai-firstify', or start a new project. Performs deep analysis across 7 dimensions, actively restructures existing projects, or guides new project setup through discovery questions. Based on the 9 design principles and 7 design patterns from the TechWolf AI-First Bootcamp.

Analyze engagement patterns across published posts to identify what works. Use when asked to review performance, find successful patterns, or optimize future content.

Generate LinkedIn post ideas from external sources (files, URLs, research). Use when the user provides source material (PDFs, URLs, articles) to brainstorm topics. NOT for writing or developing drafts - use write-linkedin-post instead.

Generate opinion piece ideas from recent LinkedIn posts (last 30 days). Use when asked to find opinion topics, brainstorm article ideas, or cross-pollinate content between LinkedIn and opinion pieces.

Entry point for the TechWolf content-studio plugin. Use to understand the workflow, pick the right content skill, or start setup for a new author/repository.

Set up a new content studio for a person. Copies the plugin template, adapts it to the person's voice, themes, and content types through interactive discovery. Use when asked to create a content studio for someone new.

Write or develop a blog post. Use for blog content - writing, drafting, developing ideas into drafts, or editing. Longer-form than LinkedIn (800-1200 words) with section structure.

Write or develop a LinkedIn post. Use ALWAYS for LinkedIn content - writing, drafting, developing ideas into drafts, or editing.

Write or develop an opinion piece (opiniestuk/op-ed). Use when asked to write opinion articles, newspaper pieces, or similar long-form opinion content.

|

|

|

|

Synthesised view of account health and activity for managers overseeing customer-facing teams (Sales, CS, Professional Services, Presales). Scans project channels, email threads, and Notion pages to surface status, risks, and upcoming milestones, without requiring the manager to trawl through individual channels. Supports proactive account management.

Comprehensive pre-meeting briefing that gathers all relevant context from Slack, email, Google Docs, Notion, and calendar. Produces a structured prep document so the manager walks into every meeting fully prepared. Supports thorough meeting preparation.

Deep-dive preparation for 1:1 meetings with direct reports. Surfaces recent work, wins, friction, wellbeing signals, and development goal progress, anchored in the org's performance framework, organizational values, and management best practices. Produces a prep sheet with suggested conversation topics, not a script.

Evidence gathering for performance review cycles. Gathers goal completion evidence, peer feedback, development progress, scope changes, and values alignment, organised along the org's performance framework dimensions, with organizational values as the 'how' lens. Surfaces evidence gaps. Never suggests ratings, only organises evidence for the manager's judgment.

Helps managers cut through noise and identify their highest-leverage actions for the day or week. Aggregates signals from calendar, triage, team context, and OKRs/goals. Presents a suggested focus list grouped by urgency, importance, and investment. The manager reviews and adjusts. Supports effective execution and prioritisation.

Interactive onboarding that discovers team structure, terminology, development goals, performance and management frameworks, organizational values, and ways of working by crawling Slack, Notion, Google Drive, Gmail, and Calendar. Validates everything with the manager before persisting. Run this first before using any other skill. Also handles periodic context refreshes via /setup --refresh.

Periodic check on team dynamics, engagement signals, and development trajectory for all direct reports. Surfaces patterns across the team: who might need more challenge, who might need more support, who hasn't had a 1:1 recently. Uses two universal lenses: performance & growth, and wellbeing & connection. Outputs are prompts for reflection, not diagnoses.

Batch-processes Slack messages and emails to surface what needs the manager's attention, categorised by urgency and type. Designed for batch-responder managers who do Slack sweeps rather than staying in reactive mode. Supports effective communication and responsiveness. Never drafts replies, only surfaces and prioritises.

Provides official TechWolf logo files in multiple variants (dark, white, monochrome) as SVG and PNG. Use when any output needs a TechWolf logo.

Build an MCP server end to end, tailored to how it will be used. Use when asked to build an MCP, create an MCP server, wrap an API as a tool, make a tool for Claude, expose a service to an agent, build a Claude connector, or turn a service into MCP tools. Asks up front who the server is for (just me, my org, or public) and what it wraps, then walks through analyze, build, deploy, scale, and distribute with steps tailored to that answer. Builds on the example-skills:mcp-builder skill for implementation depth.

Use when a task needs an isolated hidden Linux desktop or workspace-owned browser: GUI app QA, web/browser/shopping automation, sandboxed app observation, or stale workspace cleanup. Routes agent-workspace-linux MCP tools on demand. Does NOT apply to host desktop/Chrome control, generic MCP setup, or pure code/file edits.

Documentation-first development methodology. The goal is AI-ready documentation - when docs are clear enough, code generation becomes automatic. Triggers on "Build", "Create", "Implement", "Document", or "Spec out". Version 3.5 adds Phase 2.5 Adversarial Review and renames internal verification to Spec Gate (structural completeness). Clarity Gate is now a separate standalone tool for epistemic quality.

>

Analyze whether TikTok or Instagram search traffic is a viable growth channel for your business. Uses ScaleBrick's framework to evaluate demand, competition, content fit, and intent categories. Ends with a go/no-go recommendation.

Audit competitors using ScaleBrick's 3-surface framework (social, web/pages, SEO). Categorizes their pricing, features, and landing pages. Identifies gaps you can exploit, positioning angles no one is claiming, and specific moves you can make this week.

Research high-intent TikTok and Instagram search keywords using ScaleBrick's framework. Returns categorized keywords with intent type, search volume estimate, difficulty score, and content angle for each.

Generate a full marketing strategy using ScaleBrick's "TikTok as Search Engine" framework. Produces themes, pillars, voice, keyword plan, and posting schedule specific enough to execute on day one.

Inspect, monitor, and control running AI coding agent sessions across terminals via the `ctop` CLI. Use when the user asks "what agents are running", "what sessions do I have", "what is my master agent doing", "is my context about to compact", "how much have I spent", "kill the stuck session", "clean up ghost sessions", "what's idle", or anything that requires visibility across Claude Code / Codex / OpenCode sessions. Works on macOS, Linux, Windows. No network calls — reads local process state and session files.

模拟高三语文辅导老师，辅导现代文阅读、古诗文鉴赏、文言文翻译、作文写作等语文问题。重语感培养、文本解读、写作思维。当学生提出语文问题、请求分析课文、讲解古诗词、修改作文时使用。

模拟高三英语辅导老师，辅导英语阅读理解、完形填空、语法填空、写作等问题。重语言能力培养、做题技巧、词汇积累。当学生提出英语问题、请求讲解语法、分析阅读题、修改作文时使用。

模拟高三通用技术辅导老师，辅导技术设计、结构分析、流程图、算法、简单编程等通用技术问题。重实践操作、设计思维、问题解决能力培养。当学生提出技术设计、结构优化、流程设计、算法问题时使用。

模拟高三文科辅导老师，用启发式教学方法辅导政治、历史、地理等文科综合问题。侧重理解、记忆、分析能力培养。当学生提出文科问题、请求讲解历史事件、地理现象、政治原理时使用。

模拟中国高三理科辅导老师，用渐进式教学方法辅导数学、物理、化学、生物等理科问题。当学生提出理科问题、请求讲解、说"不懂"、"教我"时使用。适用于高考备考、解题辅导、概念理解。