chiruu12

Connect Nudge to a coding agent (Claude Code, Codex, or Cursor) over MCP so it can drop tasks, reminders, notes, and links into Nudge from the editor. Use when the user says "connect Nudge", "set up the Nudge MCP", "add Nudge to Claude Code / Codex / Cursor", or when the Nudge capture skill reports its tools aren't available. Optional and recommended — never force it on the user.

Capture tasks, reminders, notes, and links into Nudge (a voice assistant for developers) via its MCP tools — without leaving the editor. Use when the user says "remind me to…", "add a task", "note that…", "save this link", "don't let me forget…", or wants to stash something for later while coding. Requires the Nudge MCP server (mcp__nudge__* tools) to be connected.

Ultra-compressed communication mode. Cuts token usage ~75% by dropping filler, articles, and pleasantries while keeping full technical accuracy. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman.

Optimize AI agent context for maximum effectiveness using a 5-level hierarchy. Use when setting up a new project for AI-assisted development, when Codex keeps losing context, when responses are inconsistent across sessions, or when you want to reduce token waste. Also use when asked about "context", "rules files", or "how to set up Codex for this project".

Plan and execute deprecation and migration paths for APIs, dependencies, or patterns. Ensures nothing breaks during transition. Use when removing old code, upgrading dependencies, migrating between APIs, or when user mentions "deprecate", "migrate", "upgrade", "remove old", or "phase out".

Generate multiple radically different interface designs for a module using parallel sub-agents, then compare on simplicity, depth, and correctness of use. Based on "Design It Twice" from A Philosophy of Software Design. Use when user wants to design an API, explore interface options, compare module shapes, or mentions "design it twice".

Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.

Edit and improve articles by restructuring sections as a DAG respecting information dependencies, then rewriting each section for clarity and flow. Max 240 characters per paragraph. Use when user wants to edit, revise, or improve an article draft.

Debug CI failures in unfamiliar repo pipelines. Reads CI configs, fetches failure logs, classifies the failure type, and guides the user through diagnosing and fixing the issue. Use when CI fails on your PR in an OSS repo, when you don't understand a CI error, or when debugging GitHub Actions/CircleCI/other CI systems. Not for setting up CI from scratch — this is for debugging existing pipelines.

Evaluate whether an OSS repo is worth long-term investment before committing time. Assesses project health, governance, community, bus factor, and trajectory. Use when deciding whether to contribute to a specific repo, choosing between multiple repos, or evaluating a project's sustainability. Not for checking if a repo accepts contributions — use oss-find-issue for that.

Handle PR review feedback - understand reviewer comments, research their concerns, and help the user address them. The LLM explains what reviewers want and why; the user decides how to respond and writes the code. Use when you receive review comments on a submitted PR.

Review other contributors' PRs to learn the codebase and build maintainer trust. Guides structured diff reading, helps formulate review comments, and teaches how to distinguish blocking issues from nits. Use when you want to learn a codebase by reviewing PRs, build trust with maintainers before contributing code, or practice code review on OSS repos. Not for reviewing your own PR — use oss-submit-pr for that.

Plan and execute your second and third contributions to become a regular contributor. Covers relationship building, progressive complexity escalation, and community presence. Use when your first PR was merged and you want to keep contributing, when transitioning from one-off contributor to regular, or when planning a sustained OSS involvement. Not for your first contribution — use the oss-find-issue → oss-contribute chain for that.

Set up a complete development environment for an unfamiliar OSS repo. Handles dependency installation, building, running tests, and configuring dev tooling. Use when cloning a new OSS repo and setting up for development, when the repo's setup instructions don't work, or when you need to debug environment issues. Not for setting up your own project — this is for getting someone else's repo running.

Contribute documentation improvements to an OSS repo. Identifies doc gaps, studies the docs system, and verifies accuracy against source code. Use when contributing documentation to an open source project, fixing outdated setup instructions, adding missing API docs, or improving README clarity. Not for writing docs for your own project — this is for contributing docs to someone else's repo.

Write tests for untested OSS code as a standalone contribution. Identifies coverage gaps, studies the repo's test conventions, and guides the user through writing tests that match existing patterns. Use when contributing tests to an open source repo, looking for a low-risk first contribution, or improving test coverage. Not for writing tests alongside a bug fix — use oss-contribute for that.

Turn a feature or sprint goal into an ordered backlog of small, testable issues with dependencies. Hermes uses this to write sprints/sprint-0N.md.

Hermes's format for handing ONE scoped, testable issue to OpenClaw in

The non-negotiable architecture rules Hermes enforces when assigning and reviewing work — Laravel layering, tenancy, auth, and the human-merge gate.

Post GitHub Actions CI results to the

Define PulseDesk's domain entities and relationships before assigning model work, so the data model stays coherent. Source of truth is ARCHITECTURE.md.

Recipe for a new Eloquent model in PulseDesk — model + migration + factory + seeder + tenant trait. Use when introducing a domain entity.

Git discipline for the agent loop — branch per issue, small incremental commits, open a PR, never merge. Keeps the timeline genuine and reviewable.

Recipe for adding a REST endpoint to PulseDesk — route, controller, FormRequest, policy, JsonResource, and feature test. Use for every new API action.

Generates high-converting ad copy and visual prompts for Nike shoes tailored to specific platforms and audiences.

Generate or edit an image via any FAL image model (nano-banana edit, gpt-image, flux, ...), ROUTED THROUGH THE fal-proxy so it bills the Ads agent. image_urls must be public URLs (orchestrator hosts local product refs via MCP upload->presign). The recipe names the model + prompt. Use for keyframes, flat-cover transforms, product hero edits. (For the OpenAI gpt-image family specifically, create-image-gpt-image-fal also exists.)

Generate an instrumental music bed via ElevenLabs Music, ROUTED THROUGH THE elevenlabs-proxy so it bills the Ads agent. Trims any sparse intro, loudnorm, fades the tail. Prompt + length from the template recipe. Use for the music layer of any video-ad format.

Image-to-video (or text-to-video) via any FAL video model (Kling, Seedance, Veo), ROUTED THROUGH THE GooseWorks fal-proxy so the call bills the Ads agent. The template recipe names the model + params; image_url inputs must be public URLs (the orchestrator hosts local frames via MCP get_upload_url -> get_download_url). Returns the result video URL and downloads it. Use for the generative base clip of any video-ad format.

Generate a voiceover (VO) clip via ElevenLabs text-to-speech, ROUTED THROUGH THE elevenlabs-proxy so it bills the Ads agent. Voice id + script text come from the template recipe. Use for the spoken narration of VO-driven video-ad formats (cgi-app-sizzle, flat-vector-explainer, hypermotion). Never call ElevenLabs directly — the proxy attribution is required.

Shared helper that routes ALL paid media generation (FAL image/video, ElevenLabs music) through the GooseWorks proxies so every call bills the Ads agent — never a provider SDK's default host. Host-swaps the FAL queue URLs, loads the agent token from ~/.gooseworks/credentials.json, and returns the result CDN URL. Every video-ad media capability imports this; templates never call a provider directly.

Assemble a ChatGPT chat-reveal video ad from a thread + timeline JSON — one continuous Playwright recording of a ChatGPT mobile chat (user types with the iOS keyboard up → taps send → keyboard slides down + header cluster swaps in one beat → one gray loading dot → the assistant answer streams in word-by-word) crossfaded into a designed end card, with subliminal ChatGPT SFX and an optional ducked music bed. FREE assembly (Playwright + ffmpeg); the recipe supplies the per-brand thread + timeline + end-card config and gates the paid music call to its own capability. The ChatGPT sibling of render-imessage-chat. Use for the chatgpt-chat format.

Assemble an iMessage notification-cascade video ad (≈14s, 9:16) from a phone-on-desk plate + 3–5 messages — authentic Apple Messages banners composited in PIL (SF Pro text, green Messages icon, warm translucent-greige fill, soft shadow) spring in one-by-one at the BOTTOM and push the stack UP, a right-aligned Show-less/X pill rides above, the X clears the stack, then a serif end card resolves. FREE assembly (PIL + ffmpeg); the recipe supplies the per-brand plate, notifications, and end-card config and gates the paid plate-clean/music calls to their own capabilities. Use for the imessage-notification-cascade format.

Orchestrate DevCard MCP tools for GitHub profile analysis. Use when user wants to check a developer profile, compare developers, audit GitHub presence, generate a DevCard, or improve their profile. Entry point for all DevCard MCP interactions. Not for CLI-only operations — use devcard-generate skill for that.

Make a GitHub profile readable by AI agents and coding assistants. Use when user wants to improve their "agent readiness score", deploy devcard.json, create llms.txt, generate AGENTS.md, or make their profile machine-readable for tools like Claude Code, Cursor, Copilot, or Windsurf. Covers the full agent-readiness optimization workflow.

Generate DevCard outputs in various formats (SVG, JSON, terminal, markdown). Use when user wants to "generate a card", "create my devcard", "make an SVG", "show my profile card", embed a card in their README, or export developer data. Covers all output formats, themes, and enrichment options.

Run a full GitHub profile audit with actionable advice. Use when user says "audit my profile", "what should I improve", "review my GitHub", "how's my profile", or wants profile advice. Combines DevCard's audit scoring, advisor rules, and fix pipelines into a guided improvement workflow.

Persist an internal plan to context/product/plans/. Use when exiting plan mode, when the user says "save the plan", or when you notice a plan exists in .claude/plans/ that hasn't been copied to context/product/plans/ yet.

Run a retrospective with transcript analysis and log feedback. Use when the user explicitly invokes /retro, when a full plan implementation is complete, or after creating a PR or addressing code review feedback (if a retro hasn't happened yet this session).

Drive a fleet of headless coding agents through Marshal's MCP server: decompose a goal into independent tasks, run them in parallel in isolated git worktrees, review each diff, and integrate the good ones. Use when you have a multi-part coding goal to delegate to worker agents (Cursor, OpenCode, Codex, Antigravity, Claude Code) instead of doing it all yourself. The engine is mechanism; this playbook is the judgment - decomposition, prompt-writing, and merge decisions live here.

Author and run declarative YAML workflows on the Marshal fleet - reusable orchestration recipes (fan a goal out across clients, collect the diffs, then merge the good ones) that you run by name instead of re-planning each time. Use when the same multi-agent shape recurs (review, compare, fix-across-clients). For a one-off decomposition, use marshal-orchestrate instead. The engine runs the recipe by sequencing safe primitives; the judgment - which recipe, and which diff to keep - stays with you.

Compare AI coding strategies on a real task with Marshal: run the same goal through several configured clients (e.g. a premium model vs a cheaper one) and get a measured cost/latency/outcome comparison. Use when you want evidence for which provider/model to route a kind of work to, or to prove what a cheaper strategy actually saved. Honest by construction - it measures, it never guesses.

Gate a candidate diff behind an independent, multi-reviewer consensus before integrating it. Spawn several biased reviewers (architect, quality, tests) that each judge the SAME diff in isolation, then apply a fixed truth table - integrate only when no reviewer rejects and at least one approves. Use when a single-pass review (marshal-orchestrate step 4) is not enough: a high-stakes merge, or when you want a quorum instead of one opinion. The engine runs the agents; this playbook is the judgment.

Reach a consensus on the APPROACH before committing a fleet to build it. Spawn several biased, independent solvers that each propose a plan for the same question, then an independent judge that converges them into one concrete plan (or narrows the question and re-runs). Use when the approach is contested - multiple valid designs, ambiguous scope, or a costly wrong turn - not for obvious work. The engine runs the agents; this playbook is the judgment. Hand the converged plan to marshal-orchestrate to build it.

OpenClaw reports coding results in three sections after every task in

Decompose a Kanban or repo goal into phased steps and wait for human approval before handing work to OpenClaw.

Post a What I Did / What's Left / What Needs Your Call update to Slack.

Triage a PR's AI review comments and CI failures. Fetches bot reviews (Greptile, CodeRabbit, Copilot, etc.), classifies each comment, and diagnoses CI failures. Use when a PR has bot review comments to address, CI failures to debug, or both. Trigger: "pr triage", "check my PR", "CI failed", "bot comments", "review comments", "triage this PR", "what's blocking my PR".