Drive a pull request all the way to merged — not just check whether it is. Use this skill whenever the user wants a PR finished, says "get this PR done", "babysit this until it merges", "see this through", "make sure PROJ-123 ships", "push this over the line", or shares a PR they want completed. Built for the common case of an agent-drafted PR that needs the last mile: resolve review comments, fix the lint/type/test failures the drafting agent missed, rebase if it won't merge cleanly, mark it ready, request the right reviewer, keep the description honest, and merge once it's approved and green. It reads the PR state deterministically to decide what's blocking, then does every mechanical fix it safely can, escalating only genuine judgment calls. Never uses admin override or auto-merge. Automates as much as possible unattended; pair with `/loop` to keep driving across CI runs and review waits.
Fetch and display a quick summary of all your active issues across every board and project. Use this skill whenever the user wants a personal kanban overview, asks what they're currently assigned to, wants to know what's on their plate, or needs a cross-project status check. Trigger on phrases like: "show my issues", "what am I working on", "what's on my plate", "board summary", "active tickets", "show my issues", "what's assigned to me", "give me a status overview", "where are my tickets", "summarise my work", or any request to see a personal view across one or more boards. Don't wait for the user to name a specific tool — if they want an overview of their current work items, this skill applies.
Turn an approved spec into a terse, well-structured implementation plan that says exactly how the work will be built. Use this skill whenever the user wants an implementation plan, a technical approach, or a breakdown of how to build something they've already specced. Also trigger when the user says things like "plan out how to build this", "how should we implement X?", "write the implementation plan", "what's the technical approach?", "break down the work on PROJ-123", "turn this spec into a plan", or "how do we actually do this?". This is the step *after* the spec — where the spec captures what and why solution-agnostically, the plan owns the how. Use it even from rough notes or a bare ticket; you can read the codebase and shape a plan, asking the user any blocking questions before you write.
Triage a GitHub pull request to decide whether it's ready for human review. Use this skill whenever the user shares a PR link, asks you to check if a PR is ready, wants to triage their review queue, or mentions a pull request that needs evaluating. Also trigger when the user says things like "should I review this?", "is this PR ready?", "check this PR for me", "triage my PRs", or gives you a GitHub PR URL. If the user is an engineering lead dealing with incoming PRs from reports, this skill decides whether the PR deserves their attention right now or whether the author needs to do more work first.
Craft and send beautifully formatted Slack messages using the Slack MCP. Use this skill whenever the user wants to send a Slack message of any kind — replying to a technical question in a thread, notifying a teammate about PRs or code changes, explaining how a system works, or any communication that should look polished and professional. If the user mentions Slack, a colleague they want to message, a thread they want to reply to, or a PR they want to share — use this skill. Don't wait for the user to ask explicitly for "Slack formatting" — if they're composing a message for Slack, this skill applies.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A collection of general-purpose LLM skills that I use in my day-to-day development workflow. Each skill is a structured prompt that helps automate common engineering tasks like writing specifications and classifying work items. The prompts can be adapted for use with any LLM-powered coding tool.
| Skill | Description |
|---|---|
| spec | Write solution-agnostic specifications from raw input and classify them as a Story, Task, Refactor, Spike, Bug, or Epic — rendered in the correct standardised format for sharing with the team. Behaviour-preserving restructuring of existing code is a Refactor; net-new internal/infrastructure work is a Task. |
| plan | Turn an approved spec into a terse, structured implementation plan that owns the how — Approach, a Changes table, a detailed Breakdown grounded in the real codebase, Files touched, and Test Strategy. Resolves shape-changing ambiguity by asking up-front, so the plan is decided rather than a wall of reasoning with open questions left in. The mechanical mirror of the spec. |
| slack-message | Craft and send beautifully formatted Slack messages using the Slack MCP. Covers technical Q&A replies and PR notifications, with markdown structure, emojis as visual signposts, and guidance on including ticket links and finding recipients. |
| workflow | Orchestrate the full software delivery lifecycle — from logging an issue through refinement, implementation, and review. Determines the current step, drives work forward autonomously, and pauses only at human approval gates. |
| my-issues | Fetch and display a quick summary of all your active issues across every board and project. Shows a flat table with project, parent, issue ID, title, status, reporter, and labels — sorted by status urgency so the most actionable work appears first. |
| pr-triage | Triage a GitHub pull request to determine whether it's ready for human review. Evaluates hard gate checks (draft status, CI, spec clarity) and weighted heuristics (prior reviews, agentic code signals, blast radius). Ready PRs trigger a Slack DM; unready ones get a friendly human-sounding comment on the PR. |
| done-done | Drive a PR all the way to merged. Evaluates six deterministic gates (code complete, checks green, ready for review, peer approved, documented, merged) in a single GitHub fetch, then orchestrates one focused sub-agent per failing gate to clear it — CI fixes, doc writing, reviewer requests, and merge run in parallel where dependencies allow. Resolves author self-review comments before marking ready for peer review. Selects reviewers intelligently from file history rather than picking arbitrarily. Never uses admin override or auto-merge — merges only when all gates genuinely pass. Pair with /loop to keep driving across CI runs and review waits. |
Each skill is a self-contained markdown prompt in skills/. You can use them directly with any LLM by copying the prompt content, or install them as a plugin for supported tools.
claude plugin add philwhiteuk/llm-skills
npx claudepluginhub philwhiteuk/skillz --plugin developNo description provided.
No description provided.
Drives a full feature-development workflow (clarify -> design -> plan -> execute -> PR -> test) over Slack via the claude-slack-bridge daemon. Ships /process-setup for one-time per-repo configuration and /process for clarification + handoff to the daemon's workflow engine.
Design fluency for frontend development. 1 skill with 23 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.
Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.