Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Adds new AI providers to claude-council, configures provider API settings, troubleshoots provider connections, and documents the provider script interface. Covers creating provider shell scripts, setting API keys, and validating connectivity. Triggers on "add provider", "new AI agent", "provider not working", "API configuration", or "extend council".
Installs and configures ktx, the open-source context layer for data agents — runs ktx setup non-interactively with hidden CLI flags, configures database connections and embeddings, installs agent integration, and verifies readiness. Use when the user asks an agent to add ktx to a project, connect data sources, install agent rules, ingest schema, or troubleshoot a local ktx install.
Periodic pass that keeps a project's verification skill and feature map honest: parallel source readers per feature, one live session driving every feature, at most one PR of proven corrections. Use for /maintain-verification-skill or "audit the verify skill".
Automatic model fallback and failover. When the primary model's request fails, times out, hits rate limits, or runs out of quota, it automatically switches to a backup model to keep service continuity. Supports intelligent model selection across multiple providers and priorities, with health monitoring, automatic retries, and error recovery mechanisms.
Add or change a capability of a Boring app - an agent, a job, a chat, a tool, a view or a page - and decide whether to expose it in boring.json. Use when a task asks for a new agent or output field, a change to an agent's prompt, rules, model or output, work that runs several agents, a chat, a tool an agent or the hub calls, or a view the hub frames. Lists the files to create, the contracts to keep, the tests to run and the feature-map entry to update.
Build AI applications with OpenAI Agents SDK - text agents, voice agents (realtime), multi-agent workflows with handoffs, tools with Zod schemas, input/output guardrails, structured outputs, and streaming. Deploy to Cloudflare Workers, Next.js, or React with human-in-the-loop patterns. Use when: building text-based agents with tools and Zod schemas, creating realtime voice agents with WebRTC/WebSocket, implementing multi-agent workflows with handoffs between specialists, setting up input/output guardrails for safety, requiring human approval for critical actions, streaming agent responses, deploying agents to Cloudflare Workers or Next.js, or troubleshooting Zod schema type errors, MCP tracing failures, infinite loops (MaxTurnsExceededError), tool call failures, schema mismatches, or voice agent handoff constraints.
Operate Orca-managed worktrees, folder contexts, terminals, repos, automations, artifacts, skill sharing, worktree comments, goals/subtasks, and Orca's embedded browser through the `orca` CLI. Use when the user says "$orca-cli", "Orca worktree", "child worktree", "spawn codex/claude in a worktree", "read/wait/send Orca terminal", "handoff" / "handover" / "give this to another agent", "Orca browser", "orca artifacts", "orca goal", or "share skills". Prefer it over raw git worktree, ad hoc PTYs, or Computer Use when Orca state is involved. Use Computer Use only when a visible window needs GUI control that a CLI, filesystem, or API cannot do.
Run multiple AI coding agent sessions in parallel using git worktrees — each agent isolated in its own worktree, working on a separate branch. Use this skill whenever the user wants to: run two or more AI agents simultaneously on different features or bugs, set up isolated agent workspaces in the same repo, push parallel branches to GitHub and open/update PRs, coordinate between concurrent agent sessions, or clean up after merging. Triggers on: "parallel agents", "multiple agent sessions", "git worktree", "run agents in parallel", "work on two things at once", "isolated agent workspace", "spin up another agent", or any request involving simultaneous AI-assisted development streams.
Creates structured plans from requirements. Generates comprehensive plans with steps, dependencies, risks, and success criteria. Coordinates with specialist agents for planning input and validates plan completeness. Uses template-renderer for formatted output.
Apply to any non-trivial work, not just bulk work: edits, migrations, analyses, checks. Build the tool that does it or proves it (codemod, script, generator, or a skill your subagents follow) instead of working by hand. The tool is the artifact a reviewer can rerun
Executes agent-enhanced council queries by spawning parallel Claude subagents that each query a provider, evaluate response quality, ask follow-up questions, and return structured insights with confidence ratings and blind spot analysis. Invoked when the --agents flag is used or when complex architectural decisions are detected.
Gate on identity questions to the user. Before any "who is X?" (or role / relationship question) reaches the user, exhaust the brain's lookup chain: think → search + page read → mounted sources → timeline/graph → web. If escalation survives the chain, ask WITH a hypothesis, never a bare unknown. Also owns the no-placeholders-at-ingest rule: pages created during bulk ingestion get their relationship/role resolved immediately.
Keep a reviewable decision trail for long-running or unattended work: a TSV log with one row per decision (what, why, evidence, result). Local by default; commit it when a reviewer needs the trail to trust the result. Use for /show-me-your-work, autonomous or multi-phase runs, or work a human reviews after stepping away.
Orchestrate and coordinate coding workers in Herdr by spawning parallel Codex workers in git worktree tabs, tracking status, reconciling worker state before waiting, validating reports against raw results, processing queued reports before answering the user, correlating matching failures across streams, checking whether a constraint came from the user instruction, worker prompt, or orchestrator addition using transcript evidence, and routing task, review, and pull-request lifecycles. Use when a request mentions Herdr and asks to orchestrate, coordinate, delegate, route, reconcile state before waiting, process queued reports, correlate matching failures, explain a constraint's source, or validate worker reports, or when a prompt says you are a Herdr worker reporting to an orchestrator. Exclude single-pane or single-agent Herdr control, pane-output questions, general CLI questions, and parallel work or git worktrees without Herdr.
Author an eval for an existing skill from its REAL usage history, not its spec. Mine invocations from the brain's conversation archive (conversations/) and per-harness session transcripts — a user correction after an invocation is the gold signal — then synthesize an eval_contract plus 4-8 replayable cases with honesty labels (SPEC-DERIVED vs HISTORY-IMPLIED) and stage the result at skills/<name>/eval/autobench-<date>.md as PENDING-HUMAN-APPROVAL. Never rewrites SKILL.md. Ships two guard companions: panel integrity (multi-model judging must prove each provider actually responded) and the fail-improve taxonomy (logged LLM-fallback cases convert to deterministic code over time).
Create Claude skills from book content (markdown files). Transforms long-form book knowledge into structured, context-efficient skill packages with granular reference files, workflows, and use-case guidelines. Use this skill when: - Converting a book (markdown) into a reusable Claude skill - Creating knowledge bases from technical books, guides, or documentation - Building skills that need progressive disclosure of large content - Structuring book knowledge for efficient context loading
Before installing or using a skill, check its independent benchmark report on SkillTester.ai. Auto-trigger this skill whenever the user shows a third-party skill install command or install plan, especially `npx skills add <repo-or-url> --skill <skill-name>`, `skills add ...`, repo URL plus `--skill`, direct `SKILL.md` URLs, or `Check this skill <skill_url>`. Parse the install command, resolve the target SKILL.md, query SkillTester.ai by skill name, and return the benchmark result when the description is either an exact match or a high-overlap near match that likely represents a newer skill revision.
Use when executing I1 (implementation plan) in the sdlc-dev Spec Pack, turning requirements/design into `{FEATURE_DIR}/implementation/plan.md` (the single execution checklist and status SSOT), and providing an unambiguous task list for the subsequent I2 execution.
Execute a large multi-step project without losing the thread — track plan, progress, and decisions in a durable file. Use when the user starts a multi-day or multi-phase build, says "let's keep this organized", or work spans many sessions/handoffs.
Add knowledge bases and persistent memories to Tavus CVI personas. Use when uploading documents for RAG, enabling personas to reference PDFs/websites, persisting context across conversations, or building personas that remember users.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.