forecaster

forecaster is an agent for Claude Code from ChrisGVE/localdata-mcp. It costs 36 tokens per session (1,226 once invoked), scanned A, original, Apache-2.0.

An agent that predicts future values from measurements collected over time, such as daily sales or monthly demand. It examines trend, seasonality, stationarity, volatility, and uncertainty before choosing a forecasting method.

In plain words
What is it for?
Use it to decompose time-based data, test whether it is stable, choose ARIMA or ETS models, create forecasts, estimate uncertainty ranges, and validate the results.
Why use it?
It reduces the risk of trusting a forecast that ignores repeating patterns, changing variability, or limited historical data. It also reports uncertainty instead of presenting predictions as exact.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter.

Part of the localdata-mcp plugin — 18 skills, 11 agents, 1 MCP server shipped together

Good fit Use it to decompose time-based data, test whether it is stable, choose ARIMA or ETS models, create forecasts, estimate uncertainty ranges, and validate the results.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/chrisgve/localdata-mcp/forecaster
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/ChrisGVE/localdata-mcp

Made for: Claude Code.

Or install localdata-mcp, the plugin that ships this one along with the rest of its 18 skills, 11 agents, 1 MCP server.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for forecaster

README.md
[![agentmods](https://agentmods.dev/badge/agents/chrisgve/localdata-mcp/forecaster/github.svg)](https://agentmods.dev/agents/chrisgve/localdata-mcp/forecaster)
Your own site
<a href="https://agentmods.dev/agents/chrisgve/localdata-mcp/forecaster"><img src="https://agentmods.dev/badge/agents/chrisgve/localdata-mcp/forecaster/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for forecaster

Your own site · 80×15
<a href="https://agentmods.dev/agents/chrisgve/localdata-mcp/forecaster"><img src="https://agentmods.dev/badge/agents/chrisgve/localdata-mcp/forecaster.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 36 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,226 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00036 $0.01226
Opus 5 $0.00018 $0.00613
Sonnet 5 $0.00007 $0.00245
Haiku 4.5 $0.00004 $0.00123

Measured 8d ago against content hash 51226ddcb08e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

forecaster scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/forecaster.md · 81 lines

How it starts

The opening of the file, as written. The whole thing — 81 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a time series analysis and forecasting specialist. Your job is to decompose temporal patterns, select appropriate models based on data characteristics, produce forecasts with honest uncertainty bounds, and validate that the model actually captures the signal in the data.

Decision Framework

Before fitting any model, characterize the series:

  1. Length. Short series (< 50 observations) limit model complexity. Very short series (< 2 full seasonal cycles) preclude seasonal modeling entirely.
  2. Frequency. Identify the observation frequency (hourly, daily, weekly, monthly). This determines which seasonal periods to test.
  3. Stationarity. Run ADF and KPSS tests. If both agree the series is non-stationary, differencing is needed. If they disagree, the series is likely trend-stationary.
  4. Seasonality. Decompose the series to check for seasonal patterns. Strong seasonality points toward ETS, which handles a seasonal component directly.
  5. Trend. Linear vs. nonlinear trend affects model choice. Damped trends are safer for long-horizon forecasts.
  6. Volatility. If variance changes over time, consider log transformation or models that handle heteroscedasticity.

Workflow

  1. Extract and inspect. Use mcp__localdata__execute_query to pull the time series data. Verify it is sorted by time, check for gaps, and note the frequency.

  2. Decompose. Call mcp__localdata__analyze_time_series to separate trend, seasonal, and residual components. This reveals the dominant patterns and guides model selection.

  3. Test stationarity. Use the stationarity tests in mcp__localdata__analyze_time_series. Report ADF and KPSS results together -- they test complementary hypotheses.

  4. Select and fit a model. mcp__localdata__forecast_time_series takes a method argument with exactly two accepted values. Anything else, including sarima and prophet, is rejected with ValueError: Unknown forecast method. Choose between them from the diagnostics above:

    • "arima" (the default): good for stationary or differenced series with clear autocorrelation structure.
    • "ets" (also accepted as "exponential_smoothing"): strong for series with trend and seasonality, especially when interpretability matters.

Read the full file on GitHub · 81 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 8d ago First seen · 81 lines · 36 tokens per session scan A 51226ddcb08e

Subscribe to this mod's changes

forecaster is an agent published in the GitHub repository ChrisGVE/localdata-mcp (4 stars, last pushed 24d ago), licensed Apache-2.0. It adds 36 tokens to every session and 1,226 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

fatal-error-check

Fast pre-review check for fatal errors in LaTeX papers. Launch BEFORE full review agents (paper-critic, domain-reviewer, referee2-reviewer). Binary PASS/FAIL verdict in 15-30 seconds. Checks compilation, placeholders, broken references, number contradictions, and section completeness. Examples: Example 1: user: "Quick…

flonat/flonat-research · 180 tokens

servicenow-admin

Autonomous ServiceNow admin agent for complex multi-step tasks: CMDB audits, incident trend analysis, batch update set reviews, cross-domain investigations. Can chain 10+ API calls without user interaction.

jschuller/mcp-server-servicenow · 45 tokens

format-validator

Checks that the paper builds cleanly and meets venue formatting requirements. Internal specialist dispatched by the papermill writer and reviewer orchestrators via Task; not intended for direct invocation.

queelius/claude-anvil · 38 tokens

method-writer

Drafts methodology, algorithm, and experimental setup sections with reproducibility in mind. Internal specialist dispatched by the papermill writer orchestrator via Task; not intended for direct invocation.

queelius/claude-anvil · 40 tokens

notebook-author

Drafts and executes a chapter's paired computational notebook against plan targets. Internal specialist dispatched by the bookwright writer orchestrator and the /bookwright:notebook command via Task; not intended for direct invocation.

queelius/claude-anvil · 46 tokens

astro-manager

Use this agent when you need to retrieve, process, or analyze astronomical and astrological data. This includes fetching ephemeris data, calculating planetary positions, determining house placements, computing progressions (secondary, solar arc, tertiary, minor), or preparing astrological data for further analysis.…

manutej/luxor-claude-marketplace · 0 tokens