AI & ML

insightfulanalytics Bundle

Automatically invoke this skill whenever the user asks to refresh a semantic model or a dataset. Can also be used to manage, optimize, troubleshoot, or configure a refresh or a refresh schedule.

insightfulanalytics Bundle

Interactive naming convention standardization for TMDL-based Power BI semantic models. Automatically invoke when the user asks to "standardize naming conventions", "fix naming conventions", "clean up model names", "apply naming standards", "audit naming", "make names human readable", "rename fields", "fix abbreviations in model", or mentions renaming measures, columns, or tables for consistency across a model.

lmaick Skill

Especialista em testes ponta a ponta (E2E) com Microsoft Playwright. Aplica abordagem mobile-first (360px a 430px), Page Object Model (POM), reutilização de estado de autenticação (storageState) e locators acessíveis e resilientes.

lmaick Skill

Especialista em decomposição de produto e escrita de issues no Linear para orquestração de IAs e humanos. Padroniza especificações com escopo cirúrgico, não-escopo explícito, critérios de aceite Gherkin e vinculação de skills do repositório agent-skills.

letsloose501 Bundle

Lint, validate and security-scan Agent Skills. Use when asked to check or audit a SKILL.md; when a skill does not fire, fires on a neighbour's work, or half-works and silently skips steps; when writing or reworking one; after a rename left a reference pointing at nothing; before committing or publishing; when a skill came from elsewhere and has to be read before it is trusted; and when the question is whether the last edit made its routing better or worse.

wycats Skill

Use when investigating unfamiliar code, tracing data flows, mapping architecture, or answering questions that require adaptive exploration — where the agent needs to follow leads, use tools interactively, and synthesize findings rather than just search for known terms.

l3rowniecakez Bundle

[standard] v26.5.16 L-SKLL | Spin up coordinated agent teams for any task. Reusable framework for TeamCreate/SendMessage/TaskList patterns. Use when user says "team-agents", "spin up a team", "use teammates", "parallel agents", "coordinate agents", "fan out", or wants multiple agents working together with coordination. Do NOT trigger for simple subagent work (use Agent tool directly) or inter-Oracle messaging (use /talk-to).

l3rowniecakez Skill

[core] v26.5.16 L-SKLL | Know ourselves — show current AI identity, model info, session stats, and Oracle philosophy. Use when user asks "who are you", "who", "who we are", "what model", or wants to check current AI identity and session context. Do NOT trigger for "what is oracle" (use /about-oracle), "philosophy" or "principles" (use /philosophy), or general project questions.

yogiraja Bundle

Investigate a question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched, docs or API facts gathered, or reading legwork delegated to a background agent.

yogiraja Skill

Garbage collection for your Claude Code configuration. Periodically scans ~/.claude (skills, memory, hooks, permissions, MCP servers, caches) for redundant, stale, orphaned, or low-value items, then walks the user through a confirm-each-deletion cleanup. Use when the user says "clean up my config", "config GC", "too many skills", "audit my setup", "my .claude is bloated", or asks for a periodic config review.

yogiraja Bundle

Implement a piece of work based on a spec, a set of tickets, or an agent brief on a triaged issue.

yogiraja Bundle

Build a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.

yogiraja Bundle

Plan a huge chunk of work, more than one agent session can hold, as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.

wycats Skill

Use when walking through code, documents, or changes with the user where both perspectives contribute — the agent sees structural patterns, the user sees design intent.

wycats Skill

Use when predicting what code looks like before reading it, anticipating where a plan will encounter friction, or forming expectations specific enough to be tested. The prepare agent crystallizes this.

wycats Skill

Use when helping the user articulate their own thinking — when they have insight they haven't fully formed, or when the task requires the user to see something the agent can point to but not resolve.

wycats Skill

Use when comparing expected outcomes against actual results, judging what a gap means, or evaluating whether a change achieved its intent. The review agent crystallizes this.

wycats Skill

Use when investigating unfamiliar territory, narrowing possibilities through targeted questions, or diagnosing a problem by updating your own model of the situation.

yogiraja Bundle

Cleans up a rough, vague, or underspecified prompt the user wrote into a clear, well-structured prompt in plain English (no XML, no tags), then hands it back to read, paste, and edit rather than executing it. Adds the four things most prompts miss, a persona, an unambiguous task, two or three examples where they help, and explicit constraints. Use whenever the user shares a prompt and wants it tightened, fixed, sharpened, polished, cleaned up, de-jargoned, made more precise, or made to produce less vague or inconsistent output, usually a short rough one (e.g. 'classify each lead as hot/warm/cold', 'review my script and tell me whats wrong', 'write a cold email to a VP'), or wants a solid reusable prompt as the takeaway from a brainstorm even if they do not say "fix this prompt." This is the default for plain-English prompt cleanup. Do NOT use when the user wants the prompt wrapped in XML or wants Claude to then run it (that is the prompt-optimizer skill), or for editing ordinary prose, emails, messages, or da

yogiraja Bundle

Visualize whether skills, rules, and agent definitions are actually followed, auto-generates scenarios at 3 prompt strictness levels, runs agents, classifies behavioral sequences, and reports compliance rates with full tool call timelines. Use when checking whether agents actually follow the skills, rules, and definitions they were given, rather than assuming they do.

yogiraja Skill

Pressure-test a real decision by running it through a curated library of mental models (inversion, second-order thinking, opportunity cost, expected value, base rates / outside view, regret minimization, reversible vs irreversible decisions, margin of safety, incentives, and more), then synthesize a decisive recommendation and the one thing that would change it. Use this whenever the user is weighing a real choice and wants rigorous thinking: "help me decide", "should I do X or Y", "I'm trying to figure out whether to...", "think this through with me", "apply mental models / first principles to this", "what's the smart way to look at this decision", or is stuck between options with real stakes. Complements the-llm-council (which stages a multi-persona debate); this skill instead applies named thinking frameworks. Trigger on genuine decisions even when the user never says "mental models" by name.

yogiraja Skill

Audits Claude Code context window consumption across agents, skills, MCP servers, and rules. Identifies bloat, redundant components, and produces prioritized token-savings recommendations. Use when the context window is filling up too fast and the agents, skills, MCP servers, or rules consuming it need to be identified.

wycats Skill

Use when working with the user on tasks requiring their situated knowledge — priorities, intent, energy, context the agent can't see. Describes the relational structure of productive human-agent collaboration.

wycats Skill

Use when observations from different surfaces appear to disagree and the agent must keep each observation in its observed form before explaining causes: UI vs behavior, logs vs state, metrics vs reports, tool output vs user-visible result.

matthewye Bundle

Post-hoc audit of autopilot execution fidelity. Analyzes agent session traces to evaluate how faithfully the autopilot workflow executed against its contract, surfacing errors, friction, and drift with traceable evidence anchors. Use when the user wants to audit an autopilot run, analyze session quality, check if autopilot did what it was supposed to, or provides a session ID from an autopilot execution.

matthewye Bundle

Complete one whole spec as a single PR: a Director drives child tickets through code-run gates and a two-axis review gate, delegating code development to a role-pinned fast-model Worker. State transitions are owned by the director CLI. Use when the user asks to run a spec end to end.

yogiraja Skill

Use when the user wants to be interviewed into a written spec instead of prompting back and forth: when they already know roughly what they want built and need it pinned down fast, when they are iterating prompt-by-prompt and want to stop, or when they say "interview me", "spec me", "ask me questions first", or "help me write the spec". Ends in a spec they approve before any building. For a fuzzy idea that still needs 2-3 approaches explored first, use superpowers:brainstorming instead.

yogiraja Skill

Convene a council of 5 advisors to attack a real decision from different angles, peer-review each other anonymously, and deliver a Chairman's verdict with a concrete next step. Use this whenever the user is wrestling with a decision and asks for the LLM Council, "convene the council," "run this by the council," "what would smart people say about X," "should I X or Y," "what's the right call on...," wants multiple sharp perspectives, or wants to stress-test a choice before acting. Trigger on decision questions even when the user does not say "council" by name, as long as a real choice with stakes is on the table.

yogiraja Bundle

Transform a raw or messy ask into an XML-structured Claude prompt (role / context / task / examples / output format / constraints), tuned to the target Claude model's documented behaviour, validate it with the user, then execute it to produce the actual deliverable. The XML output is built to be used as-is and run for you, not hand-edited. Use when the user explicitly invokes /prompt-optimizer, says "optimize this prompt", "wrap this in XML tags", "rewrite this as a Claude prompt", or "structure this and run it", or pastes a long, ambiguous, multi-requirement, multi-step ask that clearly benefits from structured framing before execution. Do NOT use when the user just wants a cleaned-up, plain-English prompt handed back to read, paste, and tweak themselves, that is the prompt-fixer skill (English hygiene, no XML, no execution). Also skip simple one-liners, greetings, trivial code edits, direct file reads, or clear conversational follow-ups.

yogiraja Skill

Design a goal-oriented agent loop, or review one for the ways loops go wrong: spinning and burning tokens, Goodhart-gaming the verifier, or running a wrong answer to completion. WRITE mode gates whether to build the loop at all, then defines a machine-decidable goal, loop type, and skeleton. REVIEW mode runs an existing loop past five failure modes plus judge independence and the keep-judgment-with-the-human red lines. Use when designing an autonomous agent loop, or when you have one and worry it will spin, cheat, or finish a wrong answer.

matthewye Bundle

Post-hoc audit of autopilot execution fidelity. Analyzes agent session traces to evaluate how faithfully the autopilot workflow executed against its contract, surfacing errors, friction, and drift with traceable evidence anchors. Use when the user wants to audit an autopilot run, analyze session quality, check if autopilot did what it was supposed to, or provides a session ID from an autopilot execution.

matthewye Bundle

Post-hoc audit of autopilot execution fidelity. Analyzes agent session traces to evaluate how faithfully the autopilot workflow executed against its contract, surfacing errors, friction, and drift with traceable evidence anchors. Use when the user wants to audit an autopilot run, analyze session quality, check if autopilot did what it was supposed to, or provides a session ID from an autopilot execution.

matthewye Bundle

Autopilot task implementer. Reads the contract (AGENT-BRIEF or issue body), follows TDD discipline, auto-diagnoses errors, and produces a structured implementation report.

matthewye Bundle

Autopilot issue resolution loop: scan, dispatch implementer and reviewer sub-agents per ready-for-agent issue, then run global meta-review. Use when processing autopilot issues from any source.

matthewye Bundle

Post-hoc audit of autopilot execution fidelity. Analyzes agent session traces to evaluate how faithfully the autopilot workflow executed against its contract, surfacing errors, friction, and drift with traceable evidence anchors. Use when the user wants to audit an autopilot run, analyze session quality, check if autopilot did what it was supposed to, or provides a session ID from an autopilot execution.

matthewye Skill

Codex autopilot-director Spec run: one spec, one PR, Workers on a pinned fast model, a code-owned state machine, and a two-axis gate that must reach absolute zero.