Skill

MCP Skill

A packaged skill (yamtrack-mcp-workflow) that provides ready-to-use workflows on top of the Yamtrack MCP server. It is meant for AI assistants that connect to Yamtrack over MCP.

Location

All skill files live in the repo under src/yamtrack_mcp/skill/:

src/yamtrack_mcp/skill/
├── SKILL.md              # this skill's instructions
├── yamtrack.skill        # skill manifest (name, version, files)
├── mcp_config.json       # stdio launch config example
├── references/           # tool schemas (mcp_tools.md)
└── scripts/              # workflow scripts
    ├── mcp_client.py
    └── workflows/
        ├── research_and_track.py
        └── batch_stats.py

The skill is not deployed with the container; it is consumed by the assistant's skill loader from the source tree.

Prerequisites

The MCP server must be reachable (Streamable HTTP at /mcp, or stdio via run_mcp). See MCP for transports and authentication.

Workflows

Full parameters, steps, and examples are in the repo at src/yamtrack_mcp/skill/references/workflows.md.

Research and Track

Searches external providers for a query and, if the item is not already tracked, starts tracking it in Planning status.

from skill.scripts.workflows.research_and_track import research_and_track
await research_and_track("Dune")

Batch Statistics

Aggregates statistics across one or more date ranges for comparison.

from skill.scripts.workflows.batch_stats import get_stats_range
stats = await get_stats_range(start_date="2025-01-01", end_date="2025-12-31")

Tools Used

The workflows call the MCP tools documented in MCP: search_media, get_details, create_entry, get_statistics, and others.