Give your AI agent the keys to your projects
Two one-page prompts that teach a connected AI agent to work your Onplana projects: a planner that turns a goal into a plan.md, a task tree with dates and owners, test cases, and downstream-surface tasks, and an autonomous agent that runs those tasks, verifies them, and keeps the board honest. Plan it, then run it. Any MCP client.
Free on every plan. Bounded by your AI token allowance.
Step 1: Plan it
From a goal to a runnable plan
The planner skill turns a one-line goal into a written plan attached to the project and a task tree another agent (or a person) can pick up and run.
Write the plan
Turn the goal into a plan.md and attach it to the project as the single source of truth.
Break it into tasks
Decompose the plan into tasks and subtasks, each with acceptance criteria and an estimate.
Schedule it
Set start and due dates, wire dependencies, add milestones, and respect the working calendar.
Assign owners
Map work to people, capacity-aware, and propose assignments where the call is yours to make.
Test cases + downstream
Add test-case subtasks for every deliverable, plus tasks for the surfaces a change ripples to.
Step 2: Run it
What the autonomous agent does
A non-blocking sweep: get through every open task once, leaving each one advanced, done, or clearly annotated. No task stalls the whole run.
Pick up the open work
List a project's open tasks, move each to In Progress, and work through them one by one.
Do it in your own tools
The agent does the actual work in its own environment, then records the result back in Onplana.
Keep the board honest
Update status as you go and leave a comment trail: what you did, what you found, what is left.
Verify before marking done
Run the tests and build, and for anything user-visible drive the app in a browser, then attach the evidence to the task.
Surface problems as Issues
Blocked? Comment and move on, never stalling the run. Hit a real problem? File an Issue and link it to the task.
Kick off either one in a line
Once a skill is installed and connected, a single instruction starts it.
Plan a project
Plan [the goal] in Onplana. Create or use the project, write a plan.md and attach it, then decompose it into tasks and subtasks with start and due dates, dependencies, and owners. Add test-case subtasks for each deliverable as you go, and tasks for any downstream surfaces it touches (docs, tests, API, migrations, analytics). Set milestones, review the tree, and post a summary.
Run a project
Loop through all the remaining open tasks in [your project] and work on them autonomously. Move each task to In Progress, do the work, and verify it (run the tests and build, and for anything user-visible test it in a browser and attach a screenshot) before marking it Done, keeping each status up to date. If you are blocked or have a question, add a comment to the task and move on to the next one. If you hit a real problem, file an Issue under the same project and link it to the task. When you have been through every task, post a summary and end the session.
How to install the skill
Same three steps for any MCP client: get a token, connect the server, install the skill file. Then kick it off with a single line (the prompts are above).
- 1
Get a connection token
In Onplana, open Settings, Agents, Connect an agent and mint a personal access token (PAT). It is scoped to the projects you choose and is revocable any time.
- 2
Connect the MCP server
Point your AI client at https://mcp.onplana.com/mcp using that token. Find your client in the cards below for the exact command.
- 3
Install the skill file
Save the downloaded skill where your client looks for skills or instructions. Install the planner, the run skill, or both, the one connection serves both.
Per client
The in-app Connect an agent screen prints the exact command for your client too.
1. Connect the MCP server
claude mcp add --transport http onplana \ https://mcp.onplana.com/mcp \ --header "Authorization: Bearer <YOUR_PAT>"
2. Install the skill
# save each skill in its own folder under ~/.claude/skills: ~/.claude/skills/onplana-project-planner/SKILL.md ~/.claude/skills/onplana-autonomous-agent/SKILL.md
1. Connect the MCP server
Settings → Connectors → Add custom connector URL: https://mcp.onplana.com/mcp Auth: Bearer <YOUR_PAT>
2. Install the skill
Open or create a Project, then paste the skill into the Project instructions (or attach the .md as knowledge).
1. Connect the MCP server
Settings → Connectors → Add MCP connector URL: https://mcp.onplana.com/mcp Auth: Bearer <YOUR_PAT>
2. Install the skill
Paste the skill into your Custom GPT instructions, or attach the downloaded .md file to the chat.
1. Connect the MCP server
# ~/.codex/config.toml
[mcp_servers.onplana]
url = "https://mcp.onplana.com/mcp"
headers = { Authorization = "Bearer <YOUR_PAT>" }2. Install the skill
Add the skill to your AGENTS.md, or reference the downloaded file at the start of your prompt.
Frequently asked
What are the Onplana agent skills?▾
Which AI clients work with them?▾
Is it free?▾
Is it safe to let an agent run autonomously?▾
What do I need to get started?▾
Autonomous, not unsupervised
The skill bakes in the guardrails: the agent stays inside the projects its token allows, never marks work done it did not finish, never talks to itself, and pauses before anything irreversible or outward-facing. Prefer a slower posture? Run it in guided mode and it waits for your go-ahead at each step.