Execute code in multiple programming languages. Use when the user asks to run, test, or debug code in Python, JavaScript, shell, or other languages.

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Run this BEFORE any package install (pip / npm / apt / brew / cargo / gem / go install) you would otherwise execute via the exec tool — including when the user asks for a deliverable that needs third-party libraries (PDF generation, image processing, scrapers, dashboards, charts, OCR, audio/video, …). It searches the open agent-skill ecosystem (skills.sh) for an installable skill that already does the job; only fall back to direct package install if find-skills returns no credible match. Also covers explicit "find a skill for X" / "how do I do X" / "is there a skill that can..." requests, and any time the user expresses interest in extending agent capabilities.

Anti-detect browser automation CLI & Skills for AI agents. Use when the user needs to interact with websites with bot detection, CAPTCHAs, or anti-bot blocks, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task that requires bypassing fingerprint checks.

Self-learning system that captures corrections during sessions and reminds users to run /reflect to update CLAUDE.md. Use when discussing learnings, corrections, or when the user mentions remembering something for future sessions.

记忆审计入口。当我主动决定审视记忆质量时，先读此文件判断应使用哪个子技能。

模式提取与失效解药分析。当发现多条记忆在讲同一个教训，或发现自己在一而再再而三地犯同样的错误时使用。

节点分解。当一个节点体积过大、或塞了多个不相关概念导致disclosure无法覆盖时使用。

可发现性审计。当disclosure写法有问题、parent放错、alias缺失、子节点过多时使用。

死数据清洗。当一条记忆读不读你的行为都不会变、感悟没有现实锚点时使用。

信念对决。当父子节点内容冲突、或两条你都认可的记忆逻辑上不能并存时使用。

Retrieve machine-readable context with tanstack libraries, tanstack doc, tanstack search-docs, tanstack create --list-add-ons --json, and --addon-details for agent-safe discovery and preflight validation.

Build and iterate custom add-ons/templates with tanstack add-on init, add-on compile, add-on dev, and tanstack create --dev-watch, including sync loop preconditions, watch-path validation, and project metadata constraints.

Scaffold a TanStack app with tanstack create using --framework, --template, --toolchain, --deployment, --add-ons, --blank, and --router-only. Covers flag compatibility, non-interactive defaults, and intent-preserving command construction.

Map tanstack ecosystem partner metadata to installable add-on ids using tanstack ecosystem --json, tanstack create --list-add-ons --json, and tanstack create --addon-details --json. Covers exclusive categories, provider options, and router-only compatibility constraints.

Apply integrations to existing projects with tanstack add, including add-on id resolution, dependency chains, option prompts, and .cta.json project metadata preconditions.

Convert files and URLs to Markdown. Supports PDF, DOCX, PPTX, XLSX, HTML, EPUB, CSV, JSON, GitHub URLs, images, audio, ZIP, and more. Use when you need to extract content from any document format.

Create ShinkaEvolve task scaffolds from a target directory and task description, producing `evaluate.py` and `initial.<ext>` (multi-language). Use when asked to set up new ShinkaEvolve tasks, evaluation harnesses, or baseline programs for ShinkaEvolve.

Run existing ShinkaEvolve tasks with the `shinka_run` CLI from a task directory (`evaluate.py` + `initial.<ext>`). Use when an agent needs to launch async evolution runs quickly with required `--results_dir`, generation count, and strict namespaced keyword overrides.

Load top-performing Shinka programs into agent context using `shinka.utils.load_programs_to_df`, and emit a compact Markdown bundle for iteration planning.