deep-research-ops

Google Workspace plugin powered by the official googleworkspace/cli (gws) Agent Skills. ~95 skills for Gmail, Drive, Calendar, Sheets, Docs, Chat, Meet, Tasks, Slides, Forms, Classroom and Admin — plus role personas and ready-made recipes — all driving the gws CLI. Vendored from upstream and auto-synced weekly. Requires the gws CLI (npm i -g @googleworkspace/cli) and a one-time OAuth setup.

96 skills

Plane Agile Ops knowledge plugin. Sprint planning, task decomposition, estimation, backlog management, velocity tracking, retrospectives, standups, intake triage, modules, epics, milestones, roadmaps, dependencies, burndown, and page publishing. Written for Plane MCP 0.3.0+ (resource tools with an action parameter, PQL filters, releases); a bootstrap skill discovers the tools across naming conventions and translates for older per-operation connectors.

16 skills · 44 commands · 2 agents · hooks

Vercel ecosystem dev plugin. Deployment, AI SDK, Edge Functions, storage, routing, performance optimization. CLI deploy troubleshooting for non-Git projects, Hobby plan fixes, standalone output handling, CMS deploy hooks. Based on official vercel-plugin v0.53.0 by Vercel Labs; telemetry is opt-in (VERCEL_PLUGIN_TELEMETRY=on).

40 skills · 4 commands · 3 agents · 1 MCP server · hooks

Day-two operations for a fleet of self-hosted OpenClaw gateway instances on one Docker host. Discovers every instance from the live Docker state instead of hard-coded paths, classifies it ok / degraded / down / alien, and drives the repair work: health and liveness reporting, provider-auth triage (expired, emptied and shadowed OAuth profiles, shared-credential token sink), config surgery with a snapshot and an executable rollback, memory and embedding repair, shared skills and plugins consolidation, secret-delivery audit by key name only, security audit, version-drift and channel-aware upgrades, and reference-instance cloning. Mutations are dry-run behind an eight-block plan, need --yes, and need a typed confirmation when irreversible. Secrets are reported as fingerprint, presence and expiry — never as values. No MCP server and no required environment variables: the one optional variable, OPENCLAW_OPS_CONFIG, is an escape hatch for the fleet-config path, and deployment specifics live in that operator-owned config outside the repository. Also carries the workspace persona guides (AGENTS.md, SOUL.md, USER.md, IDENTITY.md, standing orders, bootstrap, memory, session analysis) taken over from the retired openclaw-configurator.

22 skills · 11 commands · 2 agents · hooks

Sendpulse multi-channel marketing plugin. Manage chatbots (Telegram, WhatsApp, Instagram, Messenger, Viber, TikTok), CRM (contacts, deals, pipelines, boards, tasks), Courses (LMS) academies, courses and students, email campaigns, templates, addressbooks, and SMTP transactional email via 147 MCP tools.

11 skills · 15 commands · 2 agents · 1 MCP server

Turns what a client says into documents they can correct, a machine spec an agent can build from, and a work list somebody can pick up. Keeps the macstack/ folder of a project: macstack.json — the standardized business + technical stack specification, always English — and the six client documents it is written from. OVERVIEW says what the product is and who it is for; USER-CASES carries each case with its UX bar and an addressable acceptance list; UX-UI states what each screen shows and what must never appear on it; AUTOMATION is the trigger -> task -> workflow -> role model; HANDBOOK is how a person actually uses the thing; OPEN-QUESTIONS splits what the client owes from what the team deferred. v3 makes those six pure markdown — headings and bullet lists, nothing else. No YAML blocks, no tables, no change-log sections: the only machine markup is an HTML comment the reader never sees, pointing each entity at its place in the spec. A client can edit the document in any editor and hand it back. Around them: an immutable inbox for anything a client sends, a gated delta/rulings loop that merges it, generated requirements, architecture, test cases and index that carry every id the client documents carry, an append-only ledger with one row per edit and per client comment, tasks reconciled with the team's own tracker, and a review package that shows each statement with its own history and reads the client's answers back into the ledger. Eight commands, one job each — including one that reconciles the whole folder against the source tree in a direction you have to declare: the code is master and every document is corrected through a gate that never silently overrules an answer the client gave, or the documents are master and the gaps become tasks. Every edit is journalled, every finished task sweeps the client documents and not only the generated ones, task statuses move to what the audit actually found — closing what is built and reopening what is not — and every document

18 skills · 8 commands · 1 agent · hooks

Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications

5 skills · hooks

5 skills + 35 modes + Material Passport pipeline. Includes v3.6.7 cross-model audit gate and v3.6.8 generator-evaluator contract.

5 skills · 16 commands · 4 agents · hooks

Perform AI code review on Git diffs — supports workspace changes, branch ranges, and single commits with concurrent per-file analysis, codebase search, and deep context-aware review.

2 commands

Code intelligence powered by a knowledge graph. Provides execution flow tracing, blast radius analysis, and augmented search across your codebase.

12 skills · 1 MCP server · hooks

LLM application development with LangGraph, RAG systems, vector search, and AI agent architectures for Claude 4.6 and GPT-5.4

8 skills · 3 commands · 3 agents

Long-term semantic memory for Claude Code, powered by OpenViking

5 skills · 1 command · 1 MCP server · hooks