Skip to main content
affpapa.org
RU EN
skill CX Codex

timesfm-forecasting — Claude Code skill: forecasts and setup

This skill runs zero-shot time-series forecasts with Google TimesFM. Use it for regular-grid CSV data, quantile outputs, XReg covariates, and held-out evaluation.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓ 
npx skills add K-Dense-AI/scientific-agent-skills --skill timesfm-forecasting -a claude-code
Open on GitHub

What timesfm-forecasting does

The skill prepares temporal data, checks its regular grid, and runs forecasts with TimesFM. Its default local workflow uses the TimesFM 2.5 checkpoint and can produce quantile forecasts, use XReg covariates, and evaluate predictions on held-out data. The CSV helper is scripts/forecast_csv.py; it sorts dates and checks duplicate entries, gaps, and missing values.

Before loading weights, the workflow runs scripts/check_system.py and considers RAM, cache space, and batch size. It requires Python 3.10+, timesfm 3.0.2 with PyTorch, NumPy, and pandas. XReg adds JAX and scikit-learn. Without Python 3.10+, the installation from the skill instructions will not run. TimesFM 3.0 has a separate API, and its downloaded weights are restricted to non-commercial, non-production use.

What it helps with

  • prepare a CSV on a regular time grid
  • forecast sales from historical observations
  • generate quantile prediction intervals
  • evaluate forecasts on held-out data
  • add XReg covariates to a forecast

Example prompts for the agent

  • Forecast sales from monthly_sales.csv for horizon 12 and write the result to forecasts.csv.
  • Validate this time series, find internal gaps, and prepare it for TimesFM.
  • Compare the TimesFM forecast with a naive baseline across rolling forecast origins.

How to install timesfm-forecasting

1Skills CLI (Vercel)

npx skills add K-Dense-AI/scientific-agent-skills --skill timesfm-forecasting -a claude-code

2Manually: put the files in place

~/.claude/skills/timesfm-forecasting/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/timesfm-forecasting/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about timesfm-forecasting

How do I install timesfm-forecasting in Claude Code?

Run npx skills add K-Dense-AI/scientific-agent-skills --skill timesfm-forecasting -a claude-code. The skill will be available at ~/.claude/skills/timesfm-forecasting/SKILL.md. Then install the Python dependencies from the skill instructions.

What does timesfm-forecasting do?

It prepares regular time series and runs zero-shot forecasts with TimesFM. The skill covers quantile forecasts, XReg covariates, and evaluation on held-out data.

What does this skill require?

It requires Python 3.10+, timesfm 3.0.2 with PyTorch, NumPy, and pandas. Initial checkpoint loading needs network access and cache space; XReg requires JAX and scikit-learn.

Similar skills

SKILL.md skills catalog →
find-skills vercel-labs Discovers and installs agent skills from skills.sh directly inside your coding assistant. Handy when you need domain-specific workflows for a project. CX Codex +1 4,000,000installs frontend-design official anthropics This skill guides intentional visual design for new interfaces and redesigns. Use it when a generic generated UI is not enough. 1,025,790installs grill-me Matt Pocock grill-me interviews you about a plan or design to expose gaps and sharpen decisions. Use it before implementation starts. 1,424,476installs grill-with-docs Matt Pocock This skill interviews you about a plan or design and creates ADRs and a glossary along the way. Use it when decisions need sharper definition. 1,219,756installs improve-codebase-architecture Matt Pocock This skill scans a codebase for architectural friction, then builds a visual HTML report with refactoring candidates. Use it before a deeper architecture review. 1,163,035installs design-taste-frontend Leonxlnx This skill directs the design of public pages and work screens, then checks the result in a browser through screenshots. Use it when an interface feels generic or unfinished. CX Codex +1 678,700installs

From the same set159

scikit-bio Analyzes biological sequences, alignments, and phylogenetic trees directly in your terminal. Helpful for microbiome workflows and ecology data. CX Codex +1 1,900installs consciousness-council Runs a council of thinking archetypes over a hard choice, keeps facts apart from assumptions and values, and returns a conditional recommendation. CX Codex +1 1,900installs stable-baselines3 A skill for training reinforcement learning agents with Stable Baselines3. Supports PPO, SAC, DQN and other algorithms, works with Gymnasium custom environments and vectorized rollouts. CX Codex +1 1,900installs geopandas Guides Python geospatial workflows using GeoPandas, audits vector geometries, and verifies CRS transformations before execution. CX Codex +1 1,900installs research-lookup Collects verified academic articles and references for a manuscript or review when you need 60+ sources. Uses parallel search, extracts data from PDFs and web sources, supports deep exhaustive search and optional Perplexity. CX Codex +1 1,900installs biopython Automates computational biology workflows: parses sequence files, runs alignments, and fetches records via NCBI Entrez. CX Codex +1 2,000installs torch-geometric Guides work with PyTorch Geometric graph neural networks, from graph data setup to classification tasks. Use it when a project relies on torch_geometric. CX Codex +1 2,000installs venue-templates Prepares journal manuscripts, conference papers, research posters, and grant documents with current venue rules and bundled LaTeX scaffolds. Verifies page limits, anonymity requirements, and formatting before submission. CX Codex +1 2,000installs

Project sponsors

Companies that keep this analytics open
Gold sponsor
Nutra.Club

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor