Skip to content
Get packs

TimesFM Time-Series Forecasting

Lets an agent forecast time series with Google's TimesFM model, with quantile intervals and held-out evaluation.

Source Download skill
Repository
K-Dense-AI/scientific-agent-skills
Stars
55,780
Licence
MIT
Files
15 files

What it does

This skill teaches an AI agent how to produce forecasts for ordered, time-based numbers (such as monthly sales or temperature readings) using Google's pretrained TimesFM model, without training a model for your data. The agent is guided to check that dates form a regular grid, handle missing values carefully, produce forecasts with uncertainty ranges, and test accuracy on held-out past data. It also explains which model checkpoint to use and the different license terms that apply to each.

Great for

  • Analysts who need a quick, baseline forecast of a regular time series from a CSV
  • Developers building a forecasting pipeline who want validation and reproducibility steps
  • Researchers comparing zero-shot forecasts against naive or seasonal-naive baselines

Skip it if

  • Causal questions or estimating the effect of an intervention, because the skill states TimesFM is not a causal effect estimator (forecasts describe patterns, not cause).
  • Clinical, safety-critical or climate-policy decisions: the bundled example data has unverified provenance and the anomaly labels are illustrative, not calibrated.
  • Commercial production use of the 3.0 checkpoint, because its downloaded weights are restricted to non-commercial, non-production use.

Try asking

  • “Forecast the next 12 months of sales from my monthly_sales.csv, with 80% intervals.”
  • “Check whether my weekly server-load series has regular spacing and missing values before forecasting.”
  • “Run a rolling-origin backtest of TimesFM against a seasonal-naive baseline on my data.”
  • “Build a simple anomaly screen that flags readings outside the forecast band.”
Step-by-step guide: how to use TimesFM Time-Series ForecastingInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 15 original files, the MIT licence and an attribution note.

    timesfm-forecasting.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip timesfm-forecasting.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip timesfm-forecasting.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Pin the checkpoint revision and record it with your outputs so results can be reproduced.
  • Keep the default 2.5 checkpoint if you need Apache-2.0 weight terms; read the license notice before using 3.0 in any commercial setting.
  • Always compare TimesFM against a naive baseline and treat quantile intervals as nominal until calibrated on your own data.

What's inside

  • examples/anomaly-detection/detect_anomalies.py
  • examples/covariates-forecasting/demo_covariates.py
  • examples/global-temperature/generate_animation_data.py
  • examples/global-temperature/generate_gif.py
  • examples/global-temperature/generate_html.py
  • examples/global-temperature/README.md
  • examples/global-temperature/run_example.sh
  • examples/global-temperature/run_forecast.py
  • examples/global-temperature/temperature_anomaly.csv
  • examples/global-temperature/visualize_forecast.py
  • references/api_reference.md
  • references/data_preparation.md
  • references/examples_and_validation.md
  • references/output_and_config.md
  • SKILL.md

Redistributed unmodified from K-Dense-AI/scientific-agent-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.