Task Context

SQLite-based MCP server for reusable task contexts with artifacts and full-text search

python

1

GitHub Stars

python

Language

3 months ago

First Indexed

3 weeks ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "l0kifs-task-context-mcp": {
      "command": "uv",
      "args": [
        "run",
        "python",
        "src/task_context_mcp/main.py"
      ],
      "env": {
        "TASK_CONTEXT_MCP__DATA_DIR": "./data",
        "TASK_CONTEXT_MCP__DATABASE_URL": "sqlite:///./data/task_context.db",
        "TASK_CONTEXT_MCP__LOGGING_LEVEL": "INFO"
      }
    }
  }
}

You host a lightweight MCP server that stores reusable task contexts and artifacts to guide autonomous AI improvements across similar task types. It uses SQLite with full-text search, supports artifact lifecycles, and enables agents to learn from past patterns without handling individual task instances. This makes it easy to define reusable task types and their associated practices, rules, prompts, and learnings for consistent, scalable automation.

How to use

Use an MCP client to connect to the server and perform core operations: list active task contexts, create new ones, manage artifacts, and run full-text searches across stored learnings and best practices. You will create and reuse task contexts to enable autonomous process improvement with minimal user intervention.

How to install

Prerequisites: ensure you have Python 3.12+ and the UV package manager installed on your system.

Install and run the MCP server using the provided local development flow.

Configuration and usage notes

The server stores task contexts and related artifacts in a SQLite database. It supports creating multiple artifacts per type for each task context and performs full-text search across artifacts using FTS5.

Tools and endpoints

The MCP server exposes a set of tools that let you manage task contexts and artifacts, perform searches, and reflect on learnings to update artifacts.

Troubleshooting

If you encounter startup or runtime issues, verify you are using Python 3.12+, ensure the data directory exists, and check the log level configuration to surface helpful debug information.

Available tools

get_active_task_contexts

Retrieve all currently active task contexts along with metadata such as id, summary, description, and timestamps.

create_task_context

Create a new reusable task context with a summary and detailed description.

get_artifacts_for_task_context

Fetch all artifacts for a specific task context, with options to limit types and include archived items.

create_artifact

Create a new artifact for a task context, with support for multiple artifacts per type.

update_artifact

Update an existing artifact's summary and/or content.

archive_artifact

Archive an artifact with an optional archival reason.

search_artifacts

Perform a full-text search across all artifacts and return matching results ordered by relevance.

reflect_and_update_artifacts

Reflect on learnings from task execution and prompt updates to artifacts to drive autonomous improvements.