TimesFM Forecasting
google-research/timesfm
Forecasts any univariate time series zero-shot with Google's TimesFM model, returning point forecasts and calibrated prediction intervals without training.
Methodology for probabilistic forecasting of when and whether a future event will occur.
Install Event Forecasting
$ npx skills add pymc-labs/decision-lab --skill event-forecasting -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install pymc-labs/decision-lab event-forecasting --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/pymc-labs/decision-lab.git skills-src && mkdir -p .claude/skills && cp -r skills-src/decision-packs/event-forecaster/opencode/skills/event-forecasting .claude/skills/event-forecasting && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "event-forecasting" agent skill from https://github.com/pymc-labs/decision-lab/tree/main/decision-packs/event-forecaster/opencode/skills/event-forecasting into .claude/skills/event-forecasting/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "event-forecasting", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/pymc-labs/decision-lab/tree/main/decision-packs/event-forecaster/opencode/skills/event-forecastingType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add pymc-labs/decision-lab --skill event-forecasting -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install pymc-labs/decision-lab event-forecasting --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/pymc-labs/decision-lab.git skills-src && mkdir -p .agents/skills && cp -r skills-src/decision-packs/event-forecaster/opencode/skills/event-forecasting .agents/skills/event-forecasting && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "event-forecasting" agent skill from https://github.com/pymc-labs/decision-lab/tree/main/decision-packs/event-forecaster/opencode/skills/event-forecasting into .agents/skills/event-forecasting/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "event-forecasting", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add pymc-labs/decision-lab --skill event-forecasting -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install pymc-labs/decision-lab event-forecasting --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/pymc-labs/decision-lab.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/decision-packs/event-forecaster/opencode/skills/event-forecasting .cursor/skills/event-forecasting && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "event-forecasting" agent skill from https://github.com/pymc-labs/decision-lab/tree/main/decision-packs/event-forecaster/opencode/skills/event-forecasting into .cursor/skills/event-forecasting/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "event-forecasting", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/pymc-labs/decision-lab.git --path decision-packs/event-forecaster/opencode/skills/event-forecasting--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add pymc-labs/decision-lab --skill event-forecasting -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install pymc-labs/decision-lab event-forecasting --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/pymc-labs/decision-lab.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/decision-packs/event-forecaster/opencode/skills/event-forecasting .gemini/skills/event-forecasting && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "event-forecasting" agent skill from https://github.com/pymc-labs/decision-lab/tree/main/decision-packs/event-forecaster/opencode/skills/event-forecasting into .gemini/skills/event-forecasting/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "event-forecasting", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install pymc-labs/decision-lab event-forecastingInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add pymc-labs/decision-lab --skill event-forecasting -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/pymc-labs/decision-lab.git skills-src && mkdir -p .github/skills && cp -r skills-src/decision-packs/event-forecaster/opencode/skills/event-forecasting .github/skills/event-forecasting && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "event-forecasting" agent skill from https://github.com/pymc-labs/decision-lab/tree/main/decision-packs/event-forecaster/opencode/skills/event-forecasting into .github/skills/event-forecasting/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "event-forecasting", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add pymc-labs/decision-lab --skill event-forecasting -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install pymc-labs/decision-lab event-forecasting --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/pymc-labs/decision-lab.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/decision-packs/event-forecaster/opencode/skills/event-forecasting .opencode/skills/event-forecasting && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "event-forecasting" agent skill from https://github.com/pymc-labs/decision-lab/tree/main/decision-packs/event-forecaster/opencode/skills/event-forecasting into .opencode/skills/event-forecasting/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "event-forecasting", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
event-forecasting- GitHub stars
- 199
- Token cost
- ~2.4k tokens
- SKILL.md length
- 1,071 words
- Files
- 15 (incl. references)
- Skills in repo
- 3
- Repo updated
- First seen
- Licence
- Apache-2.0
At a glance
Methodology for probabilistic forecasting of when and whether a future event will occur.
- Works in 3 steps: What kind of question is this? → How many historical cases of this exact… → What data is available?
- Any question of the form When will X happen?
- SKILL.md covers Method selection, Ten methods in this skill, When to use which reference and Core output contract, plus 1 more section
- Runs Python scripts from its folder
What it does
Event Forecasting is an agent skill from pymc-labs/decision-lab. Methodology for probabilistic forecasting of when and whether a future event will occur. Covers Bayesian survival models, reference class reasoning, driver threshold models, leading indicator models, scenario decomposition, and causal mechanism models. Use for any question of the form "When will X happen?" or "What is the probability that Y occurs by date Z?"
Its SKILL.md is about 2.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 15 other files, including reference files (for example `references/causal_mechanism.md`, `references/continuous_driver_model.md` and `references/cure_rate_model.md`).
It sits in Data & Analytics, covering Forecasting and time series. The repository describes itself as: Run tested, autonomous agent workflows on your data for meaningful decision-making. The licence is Apache-2.0.
When your agent uses it
- Any question of the form When will X happen?
- What is the probability that Y occurs by date Z?
Example prompts
- “When will X happen?”
- “What is the probability that Y occurs by date Z?”
- “/event-forecasting”
Requirements
- Python 3
Workflow steps
3 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit a68f132. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Ships script files (Python), which the agent can run.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Event Forecasting loads about 2.4k tokens when it runs, and up to ~43k if it reads all its reference files. Until then it costs about 95 tokens; SKILL.md has 1,071 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from pymc-labs/decision-lab at commit a68f132, republished under its Apache-2.0 licence (© pymc-labs). 1,071 words, ~2,393 tokens.
Download SKILL.mdSave it as
.claude/skills/event-forecasting/SKILL.md (or your agent's skills folder). This skill also uses 14 other files; get the full folder from GitHub.- name
- event-forecasting
- description
- Methodology for probabilistic forecasting of when and whether a future event will occur. Covers Bayesian survival models, reference class reasoning, driver threshold models, leading indicator models, scenario decomposition, and causal mechanism models. Use for any question of the form "When will X happen?" or "What is the probability that Y occurs by date Z?"
Event Forecasting Analysis Skill
This skill provides methodology for estimating probability distributions over when (and whether) future events will occur. It is designed for open-ended, real-world questions where the answer is uncertain, data may be sparse, and domain knowledge matters as much as statistics.
The skill is agnostic to domain. It applies equally to geopolitical events, regulatory decisions, market regime changes, clinical endpoints, supply chain resolutions, and any other time-to-event question.
Method selection
Answer these questions in order to identify which method fits your situation.
1. What kind of question is this?
| Question form | Implication |
|---|---|
| "When will X happen?" | Time-to-event. Fit a survival curve. |
| "Will X happen by date Y?" | P(event in window). All methods can answer this. |
| "How likely is X in the next N months?" | Probability over a horizon. All methods can answer this. |
| "What factors control when X happens?" | Causal structure matters. Prioritise CausalMechanismModel. |
2. How many historical cases of this exact event exist?
| Historical cases (N) | Implication |
|---|---|
| N = 0 | No historical case. HazardModel impossible. Must use broader analogues for ReferenceClassModel. |
| N = 1 | One case. HazardModel not viable. Use as threshold / calibration anchor for ThresholdCrossingModel. Force-broaden reference class. |
| N = 2–4 | Too few for reliable survival analysis. HazardModel is very prior-dominated; only run with explicit caveat. |
| N ≥ 5 | HazardModel is viable as a primary method. |
3. What data is available?
Inspect every available data file, then answer:
| Data available | Methods unlocked |
|---|---|
| Historical durations of N ≥ 5 analogous events | HazardModel (primary) |
| Historical analogues even if durations are rough | ReferenceClassModel (always) |
| Time-series of a measurable continuous driver + ≥ 1 historical case | ContinuousDriverModel, ThresholdCrossingModel |
| Continuous driver with discrete shocks / fat-tailed increments (≥ ~100 obs) + threshold | JumpDiffusionModel |
| Time-series of relevant leading indicators (updated regularly) | IndicatorModel |
| Historical transitions / dwell times across discrete regimes (calm → crisis → resolved) | MarkovStateModel |
| Domain knowledge of decision-makers or resolution mechanisms | ScenarioDecomposition, CausalMechanismModel |
| Little data, open-ended question | ReferenceClassModel + ScenarioDecomposition |
Ten methods in this skill
| Method | Data requirement | Appropriate when |
|---|---|---|
HazardModel | Historical durations of analogous events | N ≥ 5 past events; default Weibull AFT; use discrete-time cloglog variant if parametric hazard fits poorly |
ReferenceClassModel | Historical analogues (even rough ones) | Always useful as a baseline; tolerates sparse data |
IndicatorModel | Time-series of relevant signals | Leading indicators are measurable and updated regularly |
ScenarioDecomposition | Domain knowledge + expert judgment | Question has identifiable discrete paths to resolution |
CausalMechanismModel | Structural knowledge of causal drivers | Key causal factors and direction of influence are known |
ContinuousDriverModel | Continuous driver time-series (≥ 100 obs) + computable threshold | Event operationalised as driver crossing a threshold derived from data or domain knowledge |
JumpDiffusionModel | Continuous driver time-series (≥ ~100 obs) with discrete shocks / fat tails + threshold | Event is a threshold crossing and the driver moves by sudden jumps that a Gaussian model would understate |
ThresholdCrossingModel | Driver time-series + ≥ 1 historical case with known driver level | Event triggered when driver exceeds a latent estimated threshold; works with N=1 |
MarkovStateModel | Historical transitions / dwell times across discrete regimes | Situation moves through identifiable intermediate states; want transition-rate dynamics, not a static scenario snapshot |
CureRateModel | Any (works without data) | Significant probability exists that the event will NEVER resolve; standard survival models assign zero probability to permanent non-resolution |
Each forecaster selects ONE method that best fits the data summary, local data files, and prompt context. Choose the method that is most appropriate for the available evidence — you do not need to run multiple methods. The ensemble of parallel forecasters provides coverage across methods.
Implementation backend
All ten methods use raw PyMC >= 6.0 — explicit priors, custom likelihoods (censoring, state-space, mixtures), pm.Deterministic derived quantities for forecasts and psense, Nutpie sampling with Numba backend, and pm.sample_prior_predictive / pm.sample_posterior_predictive when needed. Diagnostics use ArviZ >= 1.0 (arviz_stats, arviz_plots). See each method's reference file under references/.
Show full SKILL.md (470 more words)Show less
Special case: N=1 historical event
When only one historical case of the event exists, HazardModel is not viable and HistoricalCalibration must be skipped. Adjust method selection:
- Mandatory:
ReferenceClassModelwith a broadened reference class (analogous event types, not the exact event). Seereferences/reference_class.mdfor the broadening ladder. - If a dominant causal driver is measurable:
ThresholdCrossingModel— explicitly designed for the single-case situation. - If resolution paths are identifiable:
ScenarioDecomposition. - Do NOT run
HazardModel— report "not applicable: N=1, insufficient data". PriorSensitivitybecomes especially important; flag WARN or FAIL prominently and justify insummary.md(sensitivity is not automatically a defect — seereferences/model_checks.md).
When to use which reference
| Task | Read first |
|---|---|
| Historical durations of analogous events available | references/hazard_model.md |
| Base rate from historical analogues + Bayesian updating | references/reference_class.md |
| Leading indicator regression (relevant time series available) | references/indicator_model.md |
| Explicit scenario tree (discrete resolution paths identifiable) | references/scenario_decomposition.md |
| Structural model of causal drivers | references/causal_mechanism.md |
| Continuous driver time-series + threshold crossing | references/continuous_driver_model.md |
| Continuous driver with discrete shocks / fat tails + threshold crossing | references/jump_diffusion_model.md |
| Event triggered by a driver crossing a latent estimated threshold | references/threshold_crossing.md |
| Discrete regimes with historical transitions (calm → crisis → resolved) | references/markov_state_model.md |
| Event may never resolve (permanent non-resolution possible) | references/cure_rate_model.md |
| Model checking protocols, JSON schemas, Brier score | references/model_checks.md |
| Prior sensitivity via ArviZ psense (PyMC) | references/prior_sensitivity_psense.md |
| Output schema, convergence thresholds, agreement criteria | references/output_schema.md |
Core output contract
Every method must produce forecast.json. Full schema in references/output_schema.md. Mandatory fields:
p_event_by_horizon— P(event by date) for each horizon specified in the promptmedian_days_to_eventwithp10_days/p90_daysconvergence_status(for Bayesian:OK/MARGINAL/FAIL; analytic:N_A)
Principles that override method choice
- Never fabricate numbers. If a value is
NaN, reportNaNand explain why. Never substitute 0 or a guess. - Always report intervals. Point estimates alone are forbidden. Use 94% HDI for Bayesian; 5th/95th percentile bootstraps for analytic.
- Graceful degradation. Sparse data → wider priors and broader reference classes. Never refuse to forecast because data is thin; report wider intervals.
- Calibration over precision. A well-calibrated wide interval is always better than an overconfident narrow one.
- Causal awareness. Prefer methods that explain why the event occurs over purely statistical approaches when causal structure is identifiable. Historical patterns can fail when the causal structure changes.
- Reference class discipline. When selecting a reference class, use the narrowest class with N ≥ 5 historical cases. Document why you chose it.
- No domain-specific defaults. Do not import threshold values, percentile choices, scenario structures, or reference class compositions from other forecasting tasks. Every number must be derived from the current question and data.
- Prior sensitivity is diagnostic, not a veto. Unless the user asks for causal interpretation, run PriorSensitivity on the forecast (not every model parameter). Tier A methods use psense on deterministic
p_event_by_horizon; Tier B path-simulation methods use resampled re-simulation; Tier C uses analytic perturbation. WARN/FAIL means disclose prior dependence — especially at long horizons — not that the forecast is invalid.
© pymc-labs, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 14 other files (references) in decision-packs/event-forecaster/opencode/skills/event-forecasting of pymc-labs/decision-lab.
- SKILL.md
- references/causal_mechanism.md
- references/continuous_driver_model.md
- references/cure_rate_model.md
- references/hazard_model.md
- references/indicator_model.md
- references/jump_diffusion_model.md
- references/markov_state_model.md
- references/model_checks.md
- references/output_schema.md
- references/prior_sensitivity_psense.md
- references/reference_class.md
- references/save_predictions.py
- references/scenario_decomposition.md
- references/threshold_crossing.md
Open the folder on GitHubat commit a68f132
Compare with similar skills
Event Forecasting next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Event Forecasting this skillpymc-labs/decision-lab | 199 | — | ~2.4k | Automated safety check: Pass | Apache-2.0 | |
| TimesFM Forecastinggoogle-research/timesfm | 34k | — | ~4.7k | Automated safety check: Pass | Apache-2.0 | |
| Timesfm ForecastingzLanqing/codex-claude-academic-skills | 4.6k | 6 repos | ~7.5k | Automated safety check: Notes | Apache-2.0 | |
| StatsmodelszLanqing/codex-claude-academic-skills | 4.6k | 16 repos | ~4.9k | Automated safety check: Pass | BSD-3-Clause | |
| Alphaear Predictorninehills/skills | 281 | 2 repos | ~531 | Automated safety check: Pass | None | |
| Find Hypertable Candidatestimescale/pg-aiguide | 1.9k | 1 repos | ~2.6k | Automated safety check: Pass | Apache-2.0 |
Similar skills
Timesfm Forecasting
zLanqing/codex-claude-academic-skills
Zero-shot time series forecasting with Google's TimesFM foundation model.
Statsmodels
zLanqing/codex-claude-academic-skills
Statistical models library for Python. An agent skill from zLanqing/codex-claude-academic-skills.
Alphaear Predictor
ninehills/skills
Market prediction skill using Kronos. An agent skill from ninehills/skills.
Find Hypertable Candidates
timescale/pg-aiguide
A skill your agent uses to analyze an existing PostgreSQL database and identify which tables should be converted to Timescale/TimescaleDB hypertables.
Pensieve Search
arkohut/pensieve
Search the user's local Pensieve screenshot archive by text, app, or time range.
More from pymc-labs/decision-lab
Dlab Figure Style
pymc-labs/decision-lab
decision-lab house figure style for matplotlib. An agent skill from pymc-labs/decision-lab.
Dlab CLI
pymc-labs/decision-lab
Complete reference for decision-lab (dlab). An agent skill from pymc-labs/decision-lab.
Related
Categories
Questions about Event Forecasting
What does Event Forecasting do?
Methodology for probabilistic forecasting of when and whether a future event will occur. Event Forecasting is an agent skill from pymc-labs/decision-lab. Methodology for probabilistic forecasting of when and whether a future event will occur.
When should I use Event Forecasting?
Event Forecasting fits situations like: any question of the form When will X happen?; what is the probability that Y occurs by date Z?.
How do I install Event Forecasting in Claude Code?
Run `npx skills add pymc-labs/decision-lab --skill event-forecasting -a claude-code`. Or copy the skill folder (decision-packs/event-forecaster/opencode/skills/event-forecasting in pymc-labs/decision-lab) into .claude/skills/event-forecasting in your project. Claude Code loads it when a task matches its description.
How do I install Event Forecasting in Codex?
Run `npx skills add pymc-labs/decision-lab --skill event-forecasting -a codex`. Or copy the skill folder (decision-packs/event-forecaster/opencode/skills/event-forecasting in pymc-labs/decision-lab) into .agents/skills/event-forecasting in your project. Codex loads it when a task matches its description.
Can I use Event Forecasting in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add pymc-labs/decision-lab --skill event-forecasting -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/event-forecasting, .gemini/skills/event-forecasting, .github/skills/event-forecasting and .opencode/skills/event-forecasting in your project.
What does Event Forecasting need to run?
Going by SKILL.md and its folder, Event Forecasting needs Python for the scripts in its folder. Our summary lists: Python 3.
Does Event Forecasting access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Event Forecasting safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Event Forecasting use?
Event Forecasting is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Event Forecasting use?
About 2.4k tokens (SKILL.md is roughly 9.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 41k tokens, read only when the agent opens those files.
What are the alternatives to Event Forecasting?
Skills that share tags, products or a category with Event Forecasting: TimesFM Forecasting (google-research/timesfm, 34k stars), Timesfm Forecasting (zLanqing/codex-claude-academic-skills, 4.6k stars), Statsmodels (zLanqing/codex-claude-academic-skills, 4.6k stars) and Alphaear Predictor (ninehills/skills, 281 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Event Forecasting?
pymc-labs (a GitHub organization) maintains it in pymc-labs/decision-lab, which has 199 GitHub stars. The repository holds 3 skills in this directory. The repository was last updated on September 25, 2026.
Source: pymc-labs/decision-lab on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.