Skip to main content
timesfm-forecastingby K-DenseDataGitHub stars: 47.7k

TimesFM Forecasting skill: what it does and how to install it

Performs zero-shot time-series forecasting with Google's TimesFM, preparing regular-grid CSVs, producing quantile forecasts and evaluating held-out predictions.

Summary generated from the skill's documentation.

Install

$ npx skills add K-Dense-AI/scientific-agent-skills --skill timesfm-forecasting

Run it in a terminal. If your agent is already running, start a new session so it picks the skill up.

About this skill

What it does. Uses Google's TimesFM to validate regular-grid time-series data, load the Apache-licensed 2.5 checkpoint and produce point and quantile forecasts. It supports CSV export with metadata, optional XReg covariates and rolling-origin evaluation with baselines and interval metrics.

When to use it. For zero-shot forecasting of ordered temporal targets without task-specific training. It is not intended for causal effect estimation, clinical detection or physics-based climate modelling; TimesFM 3.0 has separate API and weight-licence requirements.

History

Repo stars

47.7kAbout +18.9k since 9 Jul 2026

Before 1 Oct 2026 the curve is estimated from public event data.

Stars are counted for the whole repository, which holds 74 skills.

Show as a table
Repo stars by day
DateRepo stars
5 Oct 202647,652
4 Oct 202647,541
3 Oct 202647,444
2 Oct 202647,351
1 Oct 202647,271
24 Sept 2026 (estimated)46,976
17 Sept 2026 (estimated)46,257
10 Sept 2026 (estimated)41,836
3 Sept 2026 (estimated)30,545
27 Aug 2026 (estimated)29,347
20 Aug 2026 (estimated)29,160
13 Aug 2026 (estimated)29,027
6 Aug 2026 (estimated)28,947
30 Jul 2026 (estimated)28,894
23 Jul 2026 (estimated)28,894
16 Jul 2026 (estimated)28,841
9 Jul 2026 (estimated)28,708

Installs

1.7k

Tracking since . A chart appears once there are 7 days of data.

Installs via skills.sh

Similar skills