Catalog
Browse and download plugins, agents, skills, tools, and workflows.
96 items
Plugins
v0.2.0
Agent Core
Specialized agents: pandora, engineer, web-researcher, office-clip, pyautogui-agent, Minds, pm. Includes test workflow. System agents (compaction, title, summary) moved to agents-default (core-preinstalled).
6 agents36 skills1 workflow
v0.4.0
Default Agents
The default set of Projectflows agents: build, plan, explore, general, product-owner, and agent-owner. Includes system agents (compaction, title, summary), core skills (manage-workflow, memory-review, requirements) and product workflows.
9 agents12 skills5 workflows
v2.3.1
AI News Publisher
A review-first workflow that researches consequential AI news and publishes user-selected, verified records.
1 workflow
v0.1.0
Browser Automation
Browser control and Playwright-based browser automation.
1 tool-group
v0.1.0
Communication
Communication tools for delegation, replies, user questions, and notifications.
1 tool-group
v0.1.0
Desktop Automation
Desktop control: mouse, keyboard, screen capture, window management, and clipboard. Also includes PyAutoGUI tools for Linux/WSLg GUI automation.
2 tool-group
v0.1.0
Filesystem
Filesystem and shell tools: read, write, list files and execute shell commands.
3 tool-group
v0.1.0
MCP Presets
Curated MCP server presets for common integrations.
v0.1.0
Skill Manager
Skill management tools: list, load, search, install, create, edit, and remove skills.
7 tools
v0.1.0
Web & Search
Web access and search tools: fetch URLs and search the web with DuckDuckGo.
1 tool-group
Individual components
Agent
agent-owner
Agent owner — the go-to authority for all agent-related decisions: registry, creation, delegation advice, and improvement
from agents-default
Agent
build
The default agent. Executes tools based on configured permissions.
from agents-default
Agent
compaction
Compacts conversation history to reduce token usage.
from agents-default
Agent
engineer
Full-stack engineering orchestrator. Runs a skill-driven delivery workflow across intake, context, isolated research/planning sessions, synthesis, implementation, verification, and reporting.
from agent-core
Agent
explore
Fast agent specialized for exploring codebases. Use when you need to quickly find files by patterns, search code for keywords, or answer questions about the codebase.
from agents-default
Agent
general
General-purpose agent for researching complex questions and executing multi-step tasks. Use this agent to execute multiple units of work in parallel.
from agents-default
Agent
office-clip
Creates, reads, edits, converts, and finalizes office-style file artifacts including documents, spreadsheets, presentations, PDFs, CSVs, and related administrative deliverables.
from agent-core
Agent
pandora
First point of contact and gatekeeper. Triage coordinator that answers simple questions directly and delegates medium-to-hard tasks to specialists. Concise, direct communication.
from agent-core
Agent
plan
Plan mode. Disallows all edit tools.
from agents-default
Agent
pm
Project Manager agent specialized in orchestrating end-to-end feature implementation through coordinated sub-sessions
from agent-core
Agent
product-owner
Owns product and software work: feature intake, requirements discovery, specification, approval, and delivery handoff to the Product Engineer. Start here for any new feature, bug fix, integration, or architectural change. Produces a confirmed, build-ready specification before any implementation begins.
from agents-default
Agent
pyautogui-agent
Worker specialist for safe PyAutoGUI-based desktop automation with the full 27-tool PyAutoGUI capability set attached. Do not delegate to any other agent.
from agent-core
Agent
summary
Summarises a session.
from agents-default
Agent
title
Generates a short title for a session.
from agents-default
Agent
web-researcher
General-purpose web research specialist for discovery, source verification, and concise synthesis from current web information.
from agent-core
Skill
agents-sdk
from agent-core
Skill
ai-news-digest
Load when you need a high-recall, source-accurate digest of the latest AI news with explicit time-window confidence.
from agent-core
Skill
document-generation
Create, edit, template, and export user-outcome documents such as reports, contracts, proposals, letters, notes, and briefs as DOCX with optional PDF finalization.
from agent-core
Skill
ecosystem-autopilot
Run recurring OpenDora ecosystem self-improvement audits and controlled experiments across agents, skills, tools, sessions, schedules, routing, and coordination; use for unattended scheduled improvement loops and overnight autopilot runs.
from agent-core
Skill
evaluate-agent
Evaluate agent behavior quality with repeatable tests, scoring, and pass/fail reports. Use for agent evals, benchmarks, and regression checks.
from agent-core
Skill
evaluate-session
Read-only retrospective on a session. Audits efficiency, identifies failed and redundant tool calls, surfaces step bloat and routing errors, and produces actionable improvement recommendations. Foundation for workflow automation — once a session has been evaluated and cleaned up, the manage-workflow skill can lift the corrected step chain into a reusable workflow.
from agent-core
Skill
evaluate-skill
Evaluate skill effectiveness and trigger quality using controlled prompts and scored assertions. Use for skill evals and skill benchmarks.
from agent-core
Skill
evaluate-tool
Evaluate tool metadata quality and selection accuracy with misuse-focused test scenarios. Use for tool evaluation, tool selection tests, and parameter-fill audits.
from agent-core
Skill
goal-shaping
Load this to turn clarified requirements into executable, loop-ready GOAL.md files under .projectflows/goals/<status>/<goal-slug>/ with durable success criteria, constraints, verification expectations, and state metadata.
from agent-core
Skill
impeccable
Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a frontend interface. Covers websites, landing pages, dashboards, product UI, app shells, components, forms, settings, onboarding, and empty states. Handles UX review, visual hierarchy, information architecture, cognitive load, accessibility, performance, responsive behavior, theming, anti-patterns, typography, fonts, spacing, layout, alignment, color, motion, micro-interactions, UX copy, error states, edge cases, i18n, and reusable design systems or tokens. Also use for bland designs that need to become bolder or more delightful, loud designs that should become quieter, live browser iteration on UI elements, or ambitious visual effects that should feel technically extraordinary. Not for backend-only or non-UI tasks.
from agent-core
Skill
internal-tool-builder
Build mobile-friendly web apps for internal tools (React + FastAPI + SQLite). Use this skill when a user requests a simple internal tool like "desk check-in", "expense tracker", or "shift scheduler". Generates runnable local apps with employee-facing and manager-facing views.
from agent-core
Skill
loop-generator
Use when a user wants to turn a recurring agent task, workflow, automation, or loop-engineering idea into an executable loop spec with trigger, objective, actions, verification, and stop conditions.
from agent-core
Skill
manage-agent
Use this skill for all agent work including personas, injections, configs, routing rules, and agent eval/experiments. Do not use for skill content or tool metadata changes.
from agent-core
Skill
manage-schedule
Load when a user or agent asks to create, list, inspect, update, pause/delete, or immediately run recurring OpenDora schedules, cron jobs, reminders, digests, audits, follow-ups, reports, or scheduled agent tasks.
from agent-core
Skill
manage-session
Manage OpenDora sessions end-to-end. Load this when you need to find, inspect, organize, create via delegation, rename, reassign, reparent, archive, close, or reactivate sessions.
from agent-core
Skill
manage-skill
Required lane for all skill work including creating or updating SKILL.md and skill.json, improving trigger descriptions, and running skill eval/benchmarks. Do not use for agent persona/config or tool metadata updates.
from agent-core
Skill
manage-tool
Required lane for tool metadata only covering tool descriptions and parameter docs that affect tool selection/fill accuracy. Do not use for agent persona/config changes or skill content changes.
from agent-core
Skill
manage-workflow
Use this skill for all workflow work — creating, reading, editing, deleting, and running workflows. Guides you through composing workflows with existing node types, including parameters, variable, prompt, structured, tool, decide, set_workdir, for_each, run_workflow, configure_session, and output nodes.
from agent-core
Skill
memory-review
Review, consolidate, and clean agent memories. Use for periodic memory maintenance or on-demand cleanup.
from agent-core
Skill
nextjs
Use when building or modifying Next.js applications, routes, rendering behavior, and deployment-safe app structure decisions.
from agent-core
Skill
pdf-processing
Process, finalize, inspect, merge, split, annotate, extract, OCR, convert, and preview PDFs as a utility layer for document, spreadsheet, and presentation workflows.
from agent-core
Skill
playwright-browser
Use when automating, testing, debugging, or inspecting the OpenDora UI or any web page via Playwright MCP. Covers all 22 browser tools including navigation, observation, interaction, scripting, and control plus safety and external-impact guidance.
from agent-core
Skill
presentation-generation
Create, edit, storyboard, and export user-outcome presentations such as pitch decks, slide summaries, training decks, and visual narratives as PPTX with optional PDF export.
from agent-core
Skill
product-built
Use to implement approved plans end-to-end: create branch, apply code changes, commit coherent increments, and prepare a traceable handoff for testing.
from agent-core
Skill
product-context
Canonical skill for durable product/project context. Load when collecting requirements before a project exists, creating, reading, aligning, or updating cascading VISION.md and CONVENTIONS.md files, or producing context handoff for setup, goal shaping, planning, implementation, or audits.
from agent-core
Skill
product-explore
Use for read-only inspection of the product workspace to confirm baseline setup, identify ownership boundaries, and map exact change surfaces before planning or coding.
from agent-core
Skill
product-plan
Use after setup/exploration to convert requirements and discovered system context into an ordered implementation plan with clear file targets, dependencies, and verification strategy.
from agent-core
Skill
product-report
Use at the end of product delivery to produce one canonical, evidence-backed summary of what was delivered, what was not, verification status, and next actions.
from agent-core
Skill
product-test
Use to validate what was actually committed: inspect branch commits/diffs against requirements and plan, run evidence-based tests, and return strict PASS/FAIL/PARTIAL verdict.
from agent-core
Skill
product-verify
Use for browser-first verification against the approved plan: validate implemented UI flows/pages with Playwright-focused evidence and strict verdict output.
from agent-core
Skill
requirements
Load this when the agent must understand or clarify a user's request before planning, routing, shaping, or building. Also load alongside product-shaping for broad app/project/product creation or improvement prompts when key product requirements are still unclear, e.g. “we need a new calendar app”.
from agent-core
Skill
shadcn
Manages shadcn components and projects — adding, searching, fixing, debugging, styling, and composing UI. Provides project context, component docs, and usage examples. Applies when working with shadcn/ui, component registries, presets, --preset codes, or any project with a components.json file. Also triggers for "shadcn init", "create an app with --preset", or "switch to --preset".
from agent-core
Skill
shadcn-ui
Use when implementing or adjusting UI components with shadcn/ui patterns, tokens, and composition.
from agent-core
Skill
spreadsheet-generation
Create, edit, calculate, import, and export user-outcome spreadsheets such as budgets, financial models, trackers, tables, and CSV/XLSX deliverables.
from agent-core
Skill
suno-auth-exploration
Use when the task is to operate Suno in an authenticated browser session to locate and download tracks; verify auth first, continue if authenticated, and stop/wait only when unauthenticated.
from agent-core
Skill
vercel
Use when preparing, validating, or troubleshooting deployments and environment behavior on Vercel.
from agent-core
Workflow
Top Impact AI News Publisher
Finds consequential AI news published today, excludes stories already published, asks Mariusz to select up to nine compact cards, researches official and corroborating sources, writes an industry-impact analysis, builds a verified story record, derives X/LinkedIn posts from that finished story linking back to the OpenVibeAI site, and safely publishes verified records.
from ai-news-publisher
Workflow
Goal Implement Attempt v1
Runs one bounded implementation attempt for an existing GOAL.md. It reads the goal, inspects project context, plans the next coherent chunk, implements, verifies, and returns evidence. It does not manage the full loop or final GOAL.md lifecycle; goal-loop-v1 owns state updates and continuation.
from agents-default
Workflow
Goal Loop v1
Goal loop workflow skeleton. Current version implements Step 0 read/parse plus Step 1 gate routing. It accepts an explicit GOAL.md path, reads the file, parses YAML frontmatter/body into structured goal state, then exits or continues based on done/blocked/max-attempts/can-start gate state.
from agents-default
Workflow
Implementation Pipeline v1
Temporary structure-test version of the implementation pipeline. It preserves the high-level flow and gates but uses dummy prompt stages instead of real delegated work so the workflow graph can be validated end-to-end before enabling each real stage one at a time.
from agents-default
Workflow
Product Implement v1
Runs a simple end-to-end product implementation chain from a prepared Product Owner implementation brief. Validates the brief, branches for new-project setup or existing-project exploration, then delegates planning, implementation, testing, review, and final reporting to Product Engineer. This v1 uses one plan and one implementation agent; it does not split work into feature tracks.
from agents-default
Workflow
Product Setup v1
Initializes or normalizes a product project using fixed workflow-owned bash for invariant safety/mechanics and constrained LLM-provided command snippets only for stack-specific scaffold, install, and smoke-check phases. The LLM decides what commands to run; the workflow decides when and where to run them.
from agents-default
Workflow
AI News Data Publisher
Discovers current AI news, builds stable candidate stories, dedupes against OpenVibeAI-data AI news files, asks Mariusz which stories to publish, writes latest/archive JSON with 7-day latest retention, validates, and commits data changes when safe.
from agent-core
Workflow
Test Configure Session Node
Minimal workflow that patches the session title and confirms subsequent prompt execution.
standalone
Workflow
Test Decide Node
Minimal workflow that deterministically routes an input to one of two Output nodes.
standalone
Workflow
Test For Each Node
Minimal workflow that loops over two fixed values, captures one prompt result per item, and returns the collected array.
standalone
Workflow
Test Output Node
Minimal workflow that accepts a message and returns it through an Output node to verify structured workflow results.
standalone
Workflow
Test Prompt Node
Minimal workflow that captures a prompt response and returns it for verification.
standalone
Workflow
Test Run Workflow Child
No-input child workflow used to isolate Run Workflow session creation and result capture.
standalone
Workflow
Test Run Workflow JSON Input
Tests whether Run Workflow accepts child parameters when input is supplied as a JSON string.
standalone
Workflow
Test Run Workflow No Input
Tests Run Workflow child session creation and result capture without child parameters.
standalone
Workflow
Test Run Workflow Node
Minimal workflow that invokes the tested Output-node child workflow and returns the captured child result.
standalone
Workflow
Test Set Workdir Node
Minimal workflow that changes the workflow working directory and verifies it with pwd.
standalone
Workflow
Test Structured Node
Minimal workflow that generates schema-validated JSON and returns it for verification.
standalone
Workflow
Test Tool Node
Minimal workflow that calls the workflow registry tool and returns its result.
standalone
Workflow
Test Variable Node
Minimal workflow that creates scalar and array variables and returns them for verification.
standalone
Tools
Agents
Create, update, delete, and list agent configurations
agent_createagent_updateagent_deleteagent_listagent_get
standalone
Tools
Automation
Python-based desktop automation using PyAutoGUI for cross-platform GUI control
pyautogui_mouse_clickpyautogui_mouse_movepyautogui_mouse_move_relativepyautogui_mouse_presspyautogui_mouse_release+22
standalone
Tools
Browser
Automate web browsers with Playwright for testing and web scraping
browserplaywright_browser_mode
standalone
Tools
Communication
Ask questions and send notifications. Agent-to-agent delegation now lives in the auto-registered agent__<id> tools (Agent Delegation group) instead of a generic delegate/reply pair.
questionnotify
standalone
Tools
Desktop
Control mouse, keyboard, screen, windows, and clipboard using native OS automation
desktop_mouse_movedesktop_mouse_clickdesktop_mouse_dragdesktop_mouse_scrolldesktop_mouse_position+14
standalone
Tools
Filesystem
Read, write, edit, and search files and directories on disk
readwritelisteditglob+3
standalone
Tools
Memory
Persistent key-value memory store for reading, writing, and deleting notes across sessions
memory_readmemory_writememory_delete
standalone
Tools
Schedule
Create, manage, and run scheduled tasks and cron jobs
schedule_listschedule_createschedule_updateschedule_deleteschedule_get+1
standalone
Tools
Sessions
Search, inspect, and manage conversation sessions and their message history
session_searchsession_getsession_analyzesession_treesession_update+2
standalone
Tools
Shell
Execute shell commands and scripts
bashbatch
standalone
Tools
Skills
Load, list, search, install, create, edit, and remove skills
skill_loadskill_listskill_searchskill_installskill_create+2
standalone
Tools
System
Internal system tools for task tracking, lessons, diagnostics, and the tool registry
todowritetodoreadloglspinvalid
standalone
Tools
Tool Registry
List, inspect, and update tools in the tool registry
tool_listtool_gettool_update
standalone
Tools
Web & Search
Fetch web pages, search the web, and query code search engines
webfetchwebsearchcodesearch
standalone
Tools
Workflows
Run, create, update, delete, and inspect automated workflows
workflow_runworkflow_createworkflow_getworkflow_listworkflow_update+3
standalone