Syncs and manages task context across multiple platforms like TickTick, Notion, and GitHub using hybrid retrieval and a unified task hierarchy.
Also known as renezander030/agentic-task-system, agentic-task-system, @reneza/ats-mcp
- Stars
- 17
- Tools
- 16
- Downloads, 30 days
- —
- Latest version
- 0.18.0
- Last commit
- Nov 11, 2026
- License
- MIT
What is Agentic Task System MCP?
Agentic Task System is an MCP server. Syncs and manages task context across multiple platforms like TickTick, Notion, and GitHub using hybrid retrieval and a unified task hierarchy. It exposes 16 tools, such as find, get_task, list_projects.
Summary generated from the project's README
About Agentic Task System MCP server
The Agentic Task System (ATS) MCP server connects AI clients like Claude Desktop or Cursor to your existing task management platforms. It functions as a bridge that exposes tools for managing tasks across different services such as TickTick, Notion, or GitHub. By using a storage-agnostic design, it works with any adapter that follows the ATS contract, allowing you to use your preferred organization method, including local Markdown files or Obsidian.
Users interact with the server through a wide array of specialized tools for task manipulation and organization. The system uses hybrid retrieval, combining keyword searches with Reciprocal Rank Fusion (RRF) and dense/sparse embeddings when available. It manages complex task structures through hierarchy tools, security controls for defining action boundaries, and an append-only ledger to record agent actions. Setup requires installing the server via npm and configuring an adapter through an environment variable or a configuration file.
What you can do with it
- 1Search across multiple platformsFind specific tasks using a combination of keyword and semantic retrieval methods.
- 2Organize complex task structuresAssign specific roles and parent-child relationships to tasks to build a structured hierarchy.
- 3Enforce task security boundariesDefine trust levels and approval requirements for specific task actions.
- 4Audit agent activityMaintain a permanent record of all actions taken by an agent during task execution.
Written from the project's README. Every use case is checked against a quote from it.
How to install Agentic Task System MCP
Formats verified 10 Nov 2026Install Agentic Task System MCP in Claude Code
$ claude mcp add --env ATS_ADAPTER=<ATS_ADAPTER> agentic-task-system -- npx -y @reneza/ats-mcpInstall Agentic Task System MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"agentic-task-system": {
"command": "npx",
"args": [
"-y",
"@reneza/ats-mcp"
],
"env": {
"ATS_ADAPTER": "<ATS_ADAPTER>"
}
}
}
}Install Agentic Task System MCP in VS Code
.vscode/mcp.json
{
"servers": {
"agentic-task-system": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@reneza/ats-mcp"
],
"env": {
"ATS_ADAPTER": "<ATS_ADAPTER>"
}
}
}
}Install Agentic Task System MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"agentic-task-system": {
"command": "npx",
"args": [
"-y",
"@reneza/ats-mcp"
],
"env": {
"ATS_ADAPTER": "<ATS_ADAPTER>"
}
}
}
}Install Agentic Task System MCP in Codex
~/.codex/config.toml
[mcp_servers.agentic-task-system] command = "npx" args = ["-y", "@reneza/ats-mcp"] [mcp_servers.agentic-task-system.env] ATS_ADAPTER = "<ATS_ADAPTER>"
Install Agentic Task System MCP in Gemini CLI
~/.gemini/settings.json
{
"mcpServers": {
"agentic-task-system": {
"command": "npx",
"args": [
"-y",
"@reneza/ats-mcp"
],
"env": {
"ATS_ADAPTER": "<ATS_ADAPTER>"
}
}
}
}Agentic Task System MCP tools 16
Detected from the READMEfindget_tasklist_projectscreate_taskupdate_tasksimilarurl_forset_task_intentpromote_explorationevaluate_task_hierarchyset_task_lifecyclecheck_task_access
| find | — | |
| get_task | — | |
| list_projects | — | |
| create_task | — | |
| update_task | — | |
| similar | — | |
| url_for | — | |
| set_task_intent | — | |
| promote_exploration | — | |
| evaluate_task_hierarchy | — | |
| set_task_lifecycle | — | |
| check_task_access | — |
Show all 16 toolsShow fewer tools
add_task_linkremove_task_linktask_graphcontext_for_task
| add_task_link | — | |
| remove_task_link | — | |
| task_graph | — | |
| context_for_task | — |
Configuration
ATS_ADAPTERRequirednpm package name of the storage adapter to load, e.g. @reneza/ats-adapter-ticktick (TickTick), @reneza/ats-adapter-notion, @reneza/ats-adapter-github, @reneza/ats-adapter-beads
| Name | Description | |
|---|---|---|
| ATS_ADAPTER | npm package name of the storage adapter to load, e.g. @reneza/ats-adapter-ticktick (TickTick), @reneza/ats-adapter-notion, @reneza/ats-adapter-github, @reneza/ats-adapter-beads | Required |
Automated checks
Ran Nov 12, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachablePass
- Open source license (OSI)Pass
- Commit or release in the last yearPass
- Package name not close to a popular packagePass
- Remote endpoint uses HTTPSNot checked
- Destructive tools are labelledNot checked
- No hidden instructions to the modelPass
- No critical advisory in the package or its dependenciesPass
README
View on GitHub@reneza/ats-mcp
Model Context Protocol server for the Agentic Task System.
It exposes the task app you already use to any MCP client (Claude Desktop, Cursor, …) as a small set of tools, backed by ATS's hybrid + RRF retrieval. It's storage-agnostic: it runs over any adapter that satisfies the ~6-method ATS contract. Generic adapters get keyword + native + RRF fusion for free; embedder-backed adapters (e.g. TickTick) get the full dense/sparse hybrid.
Tools
| Tool | What it does |
|---|---|
find |
Available adapter signals plus Core keyword retrieval fused via RRF. Adapters with embeddings add dense+sparse hybrid. Pass explain: true for a per-result breakdown. |
get_task |
Fetch one item (full body) by project id + task id. |
list_projects |
List projects / folders. |
create_task |
Create an item — the agent→human write side of the two-way bus. |
update_task |
Partial update of an existing item. |
similar |
Semantically similar items (embedder-backed adapters). |
url_for |
Paste-ready deep link back to the item in its native app. |
set_task_intent |
Store outcome, completion conditions, authority, constraints, and approval requirement. |
promote_exploration |
Create a committed execution item and link the source as evidence without copying its body. |
get_task_hierarchy / set_task_hierarchy |
Read or assign exploration/goal/project/task role and one explicit parent. |
evaluate_task_hierarchy |
Check parent support, role ordering, cycles, lifecycle validity, and active explicit conflicts. |
set_task_lifecycle |
Set active/archived/superseded state and validity windows. |
get_task_security / set_task_security |
Read or define task content trust, action/resource scope, denials, and approval boundaries. |
check_task_access |
Require a reason, evaluate one scoped request, and append an allow/deny audit record. |
add_task_link |
Add a typed relationship between two tasks. |
remove_task_link |
Remove one exact typed relationship. |
task_graph |
Traverse incoming and outgoing typed relationships. |
context_for_task |
Return valid deliberate context first, then retrieval discoveries, with provenance. |
record_action / list_actions |
Write and inspect the append-only agent action ledger. |
snapshot_task_events / poll_task_events |
Establish and poll deterministic task-state events, durably staged before checkpoint advancement. Observation only; no agent execution. |
list_pending_task_events / acknowledge_task_events |
Recover unacknowledged event envelopes and explicitly remove successfully consumed ids. |
The security tools are a decision point for cooperating MCP clients. They do not sandbox tools outside ATS. The event tools emit state changes only. The local spool stores references and hashes, not task bodies. Consumers remain responsible for intent, lifecycle, access checks, and acknowledgement after successful handling.
Install & run
npm install -g @reneza/ats-mcp @reneza/ats-adapter-ticktick
ats-mcp # stdio server
The active adapter is resolved exactly like the CLI:
ATS_ADAPTERenv var, else~/.config/ats/adapter(a single line with the adapter package name), else@reneza/ats-adapter-ticktick.
Wire it into Claude Desktop
{
"mcpServers": {
"ats": {
"command": "ats-mcp",
"env": { "ATS_ADAPTER": "@reneza/ats-adapter-ticktick" }
}
}
}
Use a different store
Point it at any ATS adapter — a folder of markdown, Obsidian, Notion, your own:
ATS_ADAPTER=@you/ats-adapter-obsidian ats-mcp
See the adapter contract — six storage methods plus authentication lifecycle hooks.
- Source: https://github.com/renezander030/agentic-task-system
- License: MIT
Agentic Task System MCP questions
How do I add Agentic Task System MCP to Claude Code?
Run claude mcp add --env ATS_ADAPTER=<ATS_ADAPTER> agentic-task-system -- npx -y @reneza/ats-mcp in your terminal. Add --scope user to make it available in every project.
How do I add Agentic Task System MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What tools does Agentic Task System MCP provide?
It exposes 16 tools, including find, get_task, list_projects, create_task, update_task.
How do I change the storage provider or adapter?
You can change the adapter by setting the ATS_ADAPTER environment variable or by editing the ~/.config/ats/adapter file.
Does this server sandbox the tools it manages?
No, the security tools provided are for decision-making within the MCP client and do not sandbox tools outside of the Agentic Task System.