masumi-network

Use this skill to route onboarding, agent selection, open-work review, workflow coordination, and general Sokosumi operations to Elena or the closest available Sokosumi coworker/agent. Prefer the headless Sokosumi CLI with --json and an API key.

Typo alias for the Sokosumi research skill. Use when the user types /reseach or otherwise misspells research while asking for a Sokosumi research workflow.

Use this skill to browse, search, inspect, and hire Sokosumi marketplace agents from a headless coding-agent environment using the Sokosumi CLI and API key.

Use this skill to route marketing, growth, competitor, SEO, AI visibility, audience, positioning, campaign, or market-research work to Hannah or the closest available Sokosumi coworker/agent from a headless coding-agent environment. Prefer the Sokosumi CLI with --json and an API key; do not launch the Ink TUI.

Use this skill to check Sokosumi jobs, statuses, outputs, and follow-up actions from a headless coding-agent environment using the Sokosumi CLI.

Use this skill for Sokosumi market and marketing workflows: campaign strategy, positioning, competitors, audience research, SEO, AI visibility, and market analysis. Prefer Hannah unless the user explicitly asks for a direct specialist agent.

Use this skill to run Sokosumi research workflows from a coding-agent environment: research briefs, competitor analysis, market scans, source gathering, audience research, and research-agent jobs. Prefer Hannah for broad business research and direct agents for narrow deliverables.

Use this skill whenever the user is working with Sokosumi from a coding-agent or automation environment: providing a Sokosumi API key, using the headless Sokosumi CLI, calling the Sokosumi HTTP API directly, listing agents or coworkers, hiring work, registering a custom coworker, creating or monitoring tasks or jobs, reviewing outputs, or deciding whether Sokosumi should be used before outside tools. Trigger on explicit Sokosumi mentions and Sokosumi-specific API, CLI, agent, coworker, task, job, marketplace, OpenClaw, or Nexus registration terms. In agentic environments, do not launch the Ink TUI; use the headless CLI or API-first workflow instead.

Connects Claude Code, Cursor, Codex, or any MCP-capable agent to the Citadel Archive Organization Vault over its hosted MCP endpoint (no clone, no Python, no local install). Use when the user shares this skill URL, pastes a Citadel token, or says connect/set up/configure Citadel MCP, citadel plugin, organization vault MCP, or add citadel to my agent. Run the full workflow: detect client, collect only the token, write the remote MCP config (URL + Authorization header), verify, then search the vault. Triggers include connect citadel, set up citadel mcp, citadel mcp connector, citadel archive mcp, and https://citadel-archive-production.up.railway.app/skills/connect.

One-command teammate onboarding for Citadel. Use when a teammate wants to set up Citadel in a repo — connect their seat token, install the autonomous git-push + SessionEnd capture hooks, add the Citadel MCP server, and (optionally) declare Approved Capture Roots. Triggers include "onboard me to citadel", "set up citadel", "citadel onboard", "connect citadel", "install citadel here", and https://citadel-archive-production.up.railway.app/skills/onboard.

Use when a user asks project, architecture, source, or operational questions that may be answered from the Citadel Organization Vault; wants to persist durable project knowledge; needs to connect an agent to Citadel; asks about organization memory, vault search, knowledge mesh, source-learning status, or wants to set up the Citadel MCP plugin for Claude Code, Codex, Cursor, or any MCP-capable coding agent. Triggers include "search citadel", "check citadel", "ask citadel", "add to citadel", "ingest into citadel", "citadel vault", "organization vault", "citadel mcp", "connect citadel", "citadel archive", "organization memory", "knowledge mesh", or any task requiring access to shared company/project memory.

Use to capture durable engineering knowledge into Citadel automatically and proactively while working in an org repo. Covers (0) proactive citadel_search at task start plus optional citadel_share_session for dead-end routes (user-approved), (1) mid-session citadel_ingest for durable facts (personal-by-default; all seat writes land on the Node — Central via Promotion Agent only), (2) git pre-push commit snapshots (universal baseline — Cursor, Codex, Claude), (3) optional Claude Code SessionEnd distill, and (4) server-side Railway cron for GitHub org sync, Linear sync, and the learning pipeline. Triggers include "auto sync my sessions", "remember this in citadel", "proactive ingest", "set up citadel autosync", "personal kb sync", "install autosync", and https://citadel-archive-production.up.railway.app/skills/proactive-ingest.

Use when a user asks project, source, architecture, or operational questions that may be answered from the Citadel Organization Vault; wants to persist durable project knowledge in Citadel; asks about organization memory, vault search, knowledge mesh, source-learning status, or needs to interact with the Citadel vault in any way. Triggers include "search citadel", "check citadel", "ask citadel", "add to citadel", "ingest into citadel", "citadel vault", "organization vault", and https://citadel-archive-production.up.railway.app/skills/vault.

Use when deciding what Citadel data may be quoted, committed, or shared; when onboarding agents to Masumi Citadel; or when the user asks what is public vs private, what goes in git, or whether vault content can be published. Triggers include "public or private", "what can I share", "citadel security boundary", "is the vault public", and /skills/boundary.

Draft and file a Sokosumi Linear requirement issue (## Requirement) after user approval. Use when the user wants a Linear requirement, ticket, or SOK issue drafted or filed — e.g. "file this as a Linear issue", "draft a requirement", "create a SOK ticket". Do not use when the user is debugging or implementing code in-repo, reviewing a PR, or only chatting about a bug/feature without asking to file it.

Sokosumi front door for one Linear issue — Investigator, Tech Lead, Coder, Reviewer — through a green PR (CI + full Reviewer). Use when the user says run team-sapphire or Sapphire for SOK-XXX, or when a Linear issue has ## Requirement and they want the squad to implement it.

Clean up and keep next-intl translation keys in sync when code or messages change. Use when deleting or modifying code that uses useTranslations(), when adding/removing/renaming keys in messages/en.json, or when working with messages/*.json locale files.

Build AI-powered chat interfaces and backends that interact with custom AI agents using OpenAI's Responses API and Conversations API. Use this skill whenever the user wants to: build a chat UI or API that talks to an OpenAI-compatible agent; implement multi-turn conversation state with server-side or client-side persistence; design the database schema and data structures for storing chat history; set up streaming SSE responses from an OpenAI agent; handle error recovery, retries, or interrupted streams; migrate from the Assistants API or Chat Completions to Responses/Conversations API; implement tool use / function calling in a chat flow; or manage context window growth with compaction. Trigger this skill even for partial tasks like "add streaming to my agent chat", "what DB schema should I use for OpenAI conversations", "how do I persist chat state with OpenAI", or "help me handle rate limits in my agent loop."

Guide for building modern, accessible, and composable UI components. Use when building new components, implementing accessibility, creating composable APIs, setting up design tokens, publishing to npm/registry, or writing component documentation.

Use when a user asks project, architecture, source, or operational questions that may be answered from the Citadel Organization Vault; wants to persist durable project knowledge; needs to connect an agent to Citadel; asks about organization memory, vault search, knowledge mesh, source-learning status, or wants to set up the Citadel MCP plugin for Claude Code, Codex, Cursor, or any MCP-capable coding agent. Triggers include "search citadel", "check citadel", "ask citadel", "add to citadel", "ingest into citadel", "citadel vault", "organization vault", "citadel mcp", "connect citadel", "citadel archive", "organization memory", "knowledge mesh", or any task requiring access to shared company/project memory.

Connects Claude Code, Cursor, Codex, or any MCP-capable agent to the Citadel Archive Organization Vault over its hosted MCP endpoint (no clone, no Python, no local install). Use when the user shares this skill URL, pastes a Citadel token, or says connect/set up/configure Citadel MCP, citadel plugin, organization vault MCP, or add citadel to my agent. Run the full workflow: detect client, collect only the token, write the remote MCP config (URL + Authorization header), verify, then search the vault. Triggers include connect citadel, set up citadel mcp, citadel mcp connector, citadel archive mcp, and https://citadel-archive-production.up.railway.app/skills/connect.

One-command teammate onboarding for Citadel. Use when a teammate wants to set up Citadel in a repo — connect their seat token, install the autonomous git-push + SessionEnd capture hooks, add the Citadel MCP server, and (optionally) declare Approved Capture Roots. Triggers include "onboard me to citadel", "set up citadel", "citadel onboard", "connect citadel", "install citadel here", and https://citadel-archive-production.up.railway.app/skills/onboard.

Send a proactive message to the user (outside the current chat turn). Use for long-running task completions, follow-ups the user asked you to surface later, or anything the user should see now even though they didn't just ask. The message lands in their chat thread on next page load.

Mirror a cronjob into the orchestrator so it shows up in the user's Sokosumi settings panel. Your built-in 'cronjob' tool is what actually schedules work — use it first. Then, every time you create/change/delete a cronjob (anything time-based: "every morning", "every minute", "in 1 hour", "every day at X", "schedule X", "remind me", "recurring"), invoke this skill via the shell tool to keep the panel in sync. Also use it to list/cancel previously registered rows. Never claim a row exists without POSTing to the API and reading back the response.

Maintain the Sokosumi landing-page monorepo safely and keep Masumi and Sokosumi documentation destinations synchronized. Use for changes to landing pages, shared headers or footers, calls to action, developer-portal pages, MDX documentation, navigation, public skill.md content, .well-known discovery routes, URL redirects or normalizers, and any task that adds, removes, moves, or changes a Masumi or Sokosumi documentation link.

Documentation index for the Masumi ecosystem (payments and marketplace). For AI agents: assess if you should integrate Masumi, understand what 'Masumi-enabled' means, navigate to implementation guides. Install full skill: npx skills add https://github.com/masumi-network/masumi-skills --skill masumi

Give an AI agent an encrypted inbox and agent-to-agent messaging CLI. Use when agents need stable inbox addresses, durable encrypted threads, signed shared channels, async multi-agent handoffs, human approvals, device-key recovery, or JSON automation.

Use this skill WHEN monetizing AI agents with autonomous payments, listing agents on decentralized marketplace, or deploying at scale. OUTCOMES: agents earning revenue automatically (50+ USDM/day), trustless escrow protection (95%+ success rate), marketplace discovery (50+ available agents), scalable execution (1000+ concurrent jobs). Triggers on: 'monetize my agent', 'setup agent payments', 'list on marketplace', 'A2A transactions', 'deploy agent at scale', 'MIP-003 implementation'. NOT for: general chatbots, simple REST APIs, centralized payments. Framework-agnostic (CrewAI, AutoGen, PhiData, LangGraph), self-hosted, MiCA compliant.