菜单
SkillsMP 已收集 gbencke/dotfiles 中的 38 个 Skill。打开任一 Skill 可查看来源和详情。
已展示 38 / 38 个已收集 Skill。
Generates a detailed C4-model architecture document for the current repository: System Context, Container, Component, and Code (function/method) levels, plus a full call graph. Outputs Markdown, self-contained interactive HTML (force-directed SVG call graph,…
原文语言：英语
Query a pre-indexed CodeGraph knowledge graph of the current repo — symbol search, callers/callees, change impact, affected tests, and one-shot area exploration — so you read only the files that matter instead of grepping the tree. Use when asked to "explore…
原文语言：英语
Adversarial review of a change: a PR, a branch diff, or a patch file. Runs inline in a single process — blast radius, then per-lens propose → kill → judge — and writes a PR-comment-ready report plus findings.json under .gbencke/adversarial-review/reports/.…
原文语言：英语
Review many repos under every lens — one OS process per (repo × lens), all in parallel — then consolidate every findings.json into ONE cross-repo document grouped by severity then repo. Spawns subprocesses, never subagents. Trigger: /review-consolidate <repo>…
原文语言：英语
Adversarial review of an entire repository, run inline in a single process: propose → kill → judge, one lens per run. Writes a report plus findings.json under .gbencke/adversarial-review/reports/. Spawns no subagents. Trigger: /review-repo <path> [--lenses…
原文语言：英语
Adversarial review of a change: a PR, a branch diff, or a patch file. Runs inline in a single process — blast radius, then per-lens propose → kill → judge — and writes a PR-comment-ready report plus findings.json under .gbencke/adversarial-review/reports/.…
原文语言：英语
Review many repos under every lens — one OS process per (repo × lens), all in parallel — then consolidate every findings.json into ONE cross-repo document grouped by severity then repo. Spawns subprocesses, never subagents. Trigger: /review-consolidate <repo>…
原文语言：英语
Adversarial review of an entire repository, run inline in a single process: propose → kill → judge, one lens per run. Writes a report plus findings.json under .gbencke/adversarial-review/reports/. Spawns no subagents. Trigger: /review-repo <path> [--lenses…
原文语言：英语
Generate The Architect's Digest — a daily curated briefing of the most relevant stories in architecture, distributed systems, cloud, languages, tooling, AI agents, and CS research, sourced live from Hacker News, Lobsters, GitHub Trending, Hugging Face Papers,…
原文语言：英语
Search Telegram from the CLI — discover public channels by keyword or search messages within a channel. Use when the user wants to find Telegram channels, look up posts/messages in a channel, or do Telegram OSINT from the terminal.
原文语言：英语
Rewrites text into ASD-STE100 Simplified Technical English (STE): a controlled language with approved words, one meaning per word, active voice, short sentences, and strict rules for procedures and safety instructions. Use when user says "simplify to STE",…
原文语言：英语
Rewrites text using William Zinsser's "On Writing Well" method. Strips clutter, enforces clarity, brevity, simplicity, and humanity. Use when user says "zinsser this", "apply zinsser", "clean up my writing", "rewrite with zinsser", or invokes /zinsser.
原文语言：英语
Reorganize a long, scattered Markdown article into a clean hierarchy of Parts and Chapters, then regenerate its PDF with Pandoc (--top-level-division=part). Use when asked to "reorganize an article", "restructure into parts and chapters", "fix article…
原文语言：英语
Create a telescopic HTML study guide from any book, PDF, epub, paper, or article — skimmable key ideas that expand into depth, an interactive ideas-dependency graph, a chapter map, and deep links back to the source. Use when the user asks to "telestudy"…
原文语言：英语
Manage Guilherme Bencke's personal action-item log at 00.Tasks/ACTION_ITEMS.guilherme_bencke.md. Supports adding new items under today's date (or any explicit date), showing recent items, and searching by keyword across the full history. Use when asked to…
原文语言：英语
Run the daily workstream update workflow for the Obsidian vault. Generates today's workstream summary from meeting summaries, syncs individual workstream files, copies daily-note entries into workstream files, updates PROJECT_DECISIONS.md,…
原文语言：英语
Extract architectural decisions from an Obsidian note and render them as an inline SVG diagram — black background, white boxes and arrows, chronological top-down flow. Each box shows what was decided, what was rejected, and the meeting where the decision was…
原文语言：英语
Generate a deep-dive technical document from meeting transcripts and summaries in the Obsidian vault. Follows a strict 16-section structure with attribution rules, table preference, verbatim-only quotes, and blocker flagging. Output is written to…
原文语言：英语
Generate a per-person action items report from meeting summaries in 02.Meetings/summaries/. Groups every action item by owner across all summaries (or a date range) and writes the result as a markdown file at the repo root. Use when asked to "generate action…
原文语言：英语
Convert any *.txt transcripts at the repo root into Obsidian meeting transcript + summary pairs (offline, no API key required), then commit and push. The model authors the summary in-session from the transcript text. Use when asked to "process transcripts",…
原文语言：英语
Manage PROJECT_DECISIONS.md — a structured log of decisions made during meetings, extracted from 02.Meetings/summaries/ using the Anthropic API. File is organised by date (newest first), with categories as sub-sections under each date: architectural,…
原文语言：英语
Execute a deep research and writing pipeline using the STORM methodology, leveraging web tools and extensive human-in-the-loop feedback to produce grounded, comprehensive articles.
原文语言：英语
Apply Go monorepo and hexagonal architecture decisions for 10–50 service codebases. Covers module layout, DDD package structure, dependency rules, linting boundaries, testing, CI/CD, and modern patterns. Triggers when user asks about "Go monorepo…
原文语言：英语
Apply the Go code review checklist during PR review, design review, or self-review. Covers style & idioms, error handling, design patterns, project architecture, concurrency, testing, performance, and security. Each item tagged MUST, SHOULD, or CONSIDER.…
原文语言：英语
Detect change amplification — Ousterhout's clearest signal of bad architecture — by fetching the last N merged PRs from the current git repo and analysing them for conceptual changes that forced edits across scattered files because no module owned the…
原文语言：英语
Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or…
原文语言：英语
Structured exploration of what is possible within a given problem, decision, or system — before evaluating what is optimal. Use this skill whenever the user asks to explore options, brainstorm, map a solution space, think through alternatives, analyze what…
原文语言：英语
Run an extremely strict maintainability review for abstraction quality, giant files, and spaghetti-condition growth. Use for a thermo-nuclear code quality review, thermonuclear review, deep code quality audit, or especially harsh maintainability review.
原文语言：英语
Tell the agent to zoom out and give broader context or a higher-level perspective. Use when you're unfamiliar with a section of code or need to understand how it fits into the bigger picture.
原文语言：英语
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for…
原文语言：英语
Manage bookmarks, collections, tags, and highlights in Raindrop.io using the REST API.
原文语言：英语
Render Codex, Claude Code, OpenClaw, or Pi session JSONL transcripts as a searchable, shareable single-file HTML viewer.
原文语言：英语
Download audio and video from YouTube and 1000+ other platforms using yt-dlp.
原文语言：英语
Rewrites text using William Zinsser's "On Writing Well" method. Strips clutter, enforces clarity, brevity, simplicity, and humanity. Use when user says "zinsser this", "apply zinsser", "clean up my writing", "rewrite with zinsser", or invokes /zinsser.
原文语言：英语
Search the Context7 library index and fetch high-quality, version-accurate documentation snippets.
原文语言：英语
Structural code analysis, navigation, and diagnostics using ast-grep.
原文语言：英语
Web scraping and automation using the Obscura headless browser.
原文语言：英语
Perform AI-powered web searches to find documentation, troubleshoot errors, and summarize information.
原文语言：英语