AI & ML

nvidia-nemo-rl Skill

Configuration conventions for NeMo-RL. YAML is the single source of truth for defaults. Covers BaseModel/TypedDict usage, dataclass for internal classes, exemplar YAML updates, and forbidden default patterns.

nvidia-nemo-rl Skill

Code style guidelines for NeMo-RL (Python and shell). Covers naming, indentation, comments, docstrings, reflection avoidance, and uv usage.

technickai Skill

Use when finding new AI agent skills, discovering capabilities, installing skills from GitHub, searching skill marketplaces, or expanding what Claude can do - like Neo downloading martial arts in The Matrix

technickai Skill

Use when testing MCP servers, debugging MCP tool responses, exploring MCP capabilities, or diagnosing why an MCP tool returns unexpected data

duthaho Bundle

Interactive setup wizard for claudekit. Scaffolds rules, hooks, and MCP server configs into the user's project. Run /claudekit:init to configure. Use when setting up a new project with claudekit or reconfiguring an existing one.

0xjitsu Skill

Weighted lead scoring engine that evaluates prospects using Clay MCP enrichment data across five dimensions: company fit, title seniority, tech stack overlap, social presence, and engagement signals. Triggers when the user asks to score leads, prioritize contacts, or rank prospects. Outputs a tiered classification (Hot, Warm, Cold, Skip) with a summary table and top signal per contact for fast decision-making.

0xjitsu Skill

Merge-field personalization engine that transforms outreach templates into tailored messages using Clay MCP enrichment data. Triggers when the user asks to personalize messages, write outreach emails, or customize templates with contact-specific details. Supports four tone modes and three depth levels, from basic name insertion to hyper-personalized hooks using recent achievements and mutual connections. Enforces anti-patterns to prevent generic slop.

nobrainer-tech Bundle

Use when the owner says nb-wiki, nb-add, nb-get, nb-tidy, or llm-wiki; asks to create or query a wiki, save durable knowledge, or safely audit and maintain a Markdown knowledge base across projects or sessions; do not use for transient task state.

nobrainer-tech Bundle

Use when the owner says nb-browser or nobrainer-browser, asks to inspect or operate a rendered website, attach to an existing approved Chrome/Edge session, restart an approved non-default Chromium profile for loopback CDP, reproduce a browser flow, run Playwright tests, or record and analyze a Playwright trace. Prefer a current Playwright CLI with an explicit version readback; do not set up MCP or extra browser plugins by default.

nobrainer-tech Skill

Use when the owner says nb-security or security-review, requests a threat model or security review, or when a change crosses authentication, authorization, secrets, untrusted input, sensitive data, dependency, installer, or production trust boundaries.

nobrainer-tech Bundle

Use when the owner says nb-autoimprove, deep-autoresearch, code-autoresearch, nobrainer-capture-lesson, or nobrainer-continuous-improvement; asks to improve a skill or prompt; or wants a measurable bounded baseline-variant-evaluation loop instead of one subjective rewrite.

barretstorck Skill

Use this agent when a major project step has been completed and needs to be reviewed against the original plan and coding standards. Examples: <example>Context: The user is creating a code-review agent that should be called after a logical chunk of code is written. user: "I've finished implementing the user authentication system as outlined in step 3 of our plan" assistant: "Great work! Now let me use the code-reviewer agent to review the implementation against our plan and coding standards" <commentary>Since a major project step has been completed, use the code-reviewer agent to validate the work against the plan and identify any issues.</commentary></example> <example>Context: User has completed a significant feature implementation. user: "The API endpoints for the task management system are now complete - that covers step 2 from our architecture document" assistant: "Excellent! Let me have the code-reviewer agent examine this implementation to ensure it aligns with our plan and follows best practices" <com

chiruu12 Skill

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. Not for opening the PR in the first place, which is oss-submit-pr. If the feedback is a failing CI job, use oss-debug-ci.

chiruu12 Skill

Submit a pull request following the repo's contribution guidelines. Reviews the diff, checks for common rejection reasons, and helps the user write their own PR description. The LLM reviews. the user writes. Use when implementation is complete and tests pass. Not for responding to review comments after the PR is open. Use oss-post-pr for that.

chiruu12 Skill

Research an issue deeply and present all context to the user. but the user writes the code. The LLM investigates code paths, finds relevant files, explains patterns, and identifies constraints. The user thinks through the approach and implements it. Use when ready to start working on an issue after oss-prep-to-contribute. Not for writing tests or documentation as a standalone contribution. Use oss-write-tests or oss-write-docs for those.

del-taiseiozaki Bundle

Time-boxed technical investigation/feasibility study with Codex-first multi-agent collaboration (Codex + Opus 4.6 + Agent Teams). Codex CLI is consulted in EVERY phase for question framing, feasibility analysis, and final evaluation. Phase 1: Frame the investigation question & constraints (Claude user interaction + Codex question decomposition). Phase 2: Parallel investigation (Agent Teams: Researcher [Opus external research] + Feasibility Analyst [Codex deep analysis] + optional prototype). Phase 3: Codex synthesis into go/no-go recommendation & research report. Produces a DECISION DOCUMENT, NOT an implementation plan. Use /feature after a GO decision.

del-taiseiozaki Bundle

Comprehensive onboarding for new or returning contributors. Scans repository artifacts (git history, .claude/STATE.md, project rules, skill catalog, DESIGN.md, research & library notes, checkpoints, agent-team logs) and synthesizes a GUIDE.md at the repository root summarizing what has been worked on, why, and how to resume work.

del-taiseiozaki Bundle

Unified feature planning & implementation skill — replaces the old /add-feature and /start-feature skills (both trigger phrases still apply here). MODE=existing (formerly /add-feature): add a feature to an established codebase with Codex-first collaboration — Codex is consulted in every phase for scope analysis, architecture design, implementation planning, and validation. MODE=greenfield (formerly /start-feature): start a large or new feature that requires external research — Agent Teams (Researcher + Architect) do parallel research & design. Both modes share Phase 3 complexity routing (SIMPLE: Codex direct, MODERATE: Codex + /team-execute --review-only, COMPLEX: /team-execute).

del-taiseiozaki Bundle

Two-phase Agent Teams execution — replaces the old /team-implement and /team-review skills. Phase 1 IMPLEMENT (formerly /team-implement): parallel implementation with teammates per module/layer, file-ownership separation, and a shared task list with dependencies. Run after /feature plan approval. Phase 2 REVIEW (formerly /team-review): specialized reviewers (security, quality, test coverage) review the changes from different perspectives in parallel. Pass --review-only to skip Phase 1 and review existing changes (after manual or Codex implementation).

del-taiseiozaki Bundle

Diagnose and plan fixes for errors/bugs with Codex-first multi-agent collaboration (Codex + Opus 4.6 + Agent Teams). Codex CLI is consulted in EVERY phase for deep code reasoning, hypothesis evaluation, and fix validation. Phase 1: Error reproduction & context gathering (Opus subagent 1M context + Codex initial analysis + Claude user interaction). Phase 2: Parallel diagnosis (Agent Teams: Root Cause Analyst [Codex-driven] + Impact Investigator [Opus + Codex risk analysis]). Phase 3: Fix plan synthesis, Codex validation & user approval. Fix implementation is handled separately by /team-execute.

dogkeeper886 Skill

Author, review, and refactor Agent Skills — SKILL.md frontmatter, description wording for reliable triggering, progressive disclosure across bundled files, workflow checklists, validator feedback loops, and evaluation-driven iteration. Use whenever the user is writing a new skill, editing or splitting an existing SKILL.md, migrating slash commands or prompts into skills, packaging skills into a plugin, debugging a skill that fails to trigger or triggers too often, or asking how skills should be structured — including when they describe the work as "turn this into a skill" or "make this reusable" without saying the word skill.

dogkeeper886 Skill

Does the work a claimed issue describes — writes and commits the change, or performs the act and captures the evidence that it happened. Holds the scope the issue set, records why in the commit message rather than only what, and ends by stating what state the work is actually in rather than declaring it done. The write half of an issue; something else judges whether it finished.

dogkeeper886 Skill

Merges an approved change request and cleans up after it: the code lands in the default branch, the branch is gone, the issue closes, the labels clear. Asks a person outright whether the change was reviewed and tested at this exact head, records that answer, and pins the merge to that SHA so a head that moved cannot slip through. Refuses to merge without the answer.

dogkeeper886 Skill

Pushes a committed branch and opens the change request that carries it to a human, linked to its issue so the issue closes on merge rather than by hand. Moves the commits onto a branch first when they landed on the default one, gets the closure keyword right, and stops for review. It does not merge, and it does not judge whether the work is good; something before it did that.

dogkeeper886 Skill

Splits work too large for one issue into issues that can each be delivered on their own, and groups them with whatever the project already uses to group things — a milestone, a project board, an epic. Decides first whether splitting is warranted at all, since most work is not, and never invents a grouping the platform already provides.

dogkeeper886 Bundle

Turns something said in a session into a tracked issue that a cold agent can act on, and copies the session log beside it wherever one can be had — always attempted, never a reason to withhold the issue. Decides first whether the thing needs tracking at all, searches for an existing issue before creating a second one, quotes the request in the words it was made in rather than paraphrasing it, and states an acceptance test someone else can check. Creates the portal every later step resolves back to.

dogkeeper886 Skill

Writes a project's README from the code: researches current README convention on the web, studies the repo from its ground truth rather than its existing docs, finds the key points a newcomer needs first, draws every one of them as an SVG rendered to PNG, and writes the file. Produces the README only — it does not open a PR, and review-readme gates what it wrote.

dogkeeper886 Skill

Picks up a tracked issue of any kind and claims it: resolves which one, reads it whole along with the log it links, judges whether it can actually be started, and refuses it back when it cannot. Assigns it, moves the state label, cuts a branch where a diff is coming, and stops before any work begins — so a misreading is catchable while it is still cheap. The read half of an issue; something else does the work.

dogkeeper886 Skill

Judges what state finished work is actually in, and routes it there. Checks the acceptance line against the work rather than against the claim about it, sweeps what the change left behind, and sends the result onward, back for more, back to the issue, or up for splitting. Only one of its five outcomes reaches a change request, which is the point of having it.

dogkeeper886 Skill

Gates a README before a person reads it, on three counts: it delivers for a newcomer, it is true to the code, and it reads and looks right. Derives the key points from the code and checks the README leads with them and ships a diagram for each, verifies every command, path, env var and link against the source, and re-renders each diagram to catch one that has drifted from what it is drawn from. Gates what gen-readme wrote, and any README edited by hand.

dogkeeper886 Bundle

Finds and removes what this plugin's earlier versions wrote into a project: forked commands and skills that shadow the placed units, rules and story files nothing reads any more, and the sections this plugin wrote into `project-profile.md`. Shows every file before touching one, and deletes only on a yes. It never adds: every edit it makes is a removal.

dogkeeper886 Skill

Reports work in two lines and a question — the verdict, the one next step, and whether the reader wants it run or wants the detail. Findings, what was checked, risks and decisions are prepared and held; the reader acts on the next step or asks. Only a blocked or failed next step brings its blocker up with it. Replaces the reply that mixes progress, findings, reasoning and status into a narrative the reader has to mine.

dogkeeper886 Bundle

Reviews the words of a document a person will read, and returns PASS or REVISE. Starts by establishing what the file actually is and who reads it, because a markdown file carries no contract the way a source file does, and the same wording is correct in a talk track and wrong in a test case. Then runs check-prose.sh to locate the mechanical tells by grep rather than by impression, and judges what a pattern cannot reach.

dogkeeper886 Skill

Gates the tooling an agent reads — skills, commands, rules, CLAUDE.md — asking whether the file has one clear job, delivers it end to end, leaves the choices the project owns to the project, matches the repo as it actually is, and can be followed without guessing. Returns PASS, REVISE or CUT. CUT is the verdict nothing else makes: this artifact duplicates another or does nothing, and should be removed.

juniyadi Skill

This skill should be used when the user asks to "plan a feature", "evaluate a feature", "review a feature proposal", "analyze a feature before implementation", "plan a feature", "review this feature", "assess this feature", or uses the /super-dev command for multi-agent feature planning with 5-perspective judgment system.

swarm-ai-safety Bundle

SWARM: System-Wide Assessment of Risk in Multi-agent systems. Simulate multi-agent dynamics, test governance, study emergent risks.