A machine-learning explanation method based on Shapley values, which estimate how much each input feature contributed to a model's prediction. FastSHAP trains a separate explainer to produce these feature-importance results in one pass.

Use it to train explainers for tabular or image models, create surrogate models that handle masked inputs, and produce local feature-importance results in real time or batches.

It helps explain individual predictions without repeatedly running a slower explanation procedure for every example. The explanations can show which table fields or image regions influenced an output.

fastshap — Skill for Claude Code by zjunlp

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.

Any agent

npx skills add zjunlp/Mechanist --skill amortized-shap

Clone the repo

git clone --depth 1 https://github.com/zjunlp/Mechanist

Made for: Claude Code.

Or install mechanist, the plugin that ships this one along with the rest of its 54 skills, 4 agents.

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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/zjunlp/mechanist/amortized-shap/github.svg)](https://agentmods.dev/skills/zjunlp/mechanist/amortized-shap)

Your own site

<a href="https://agentmods.dev/skills/zjunlp/mechanist/amortized-shap"><img src="https://agentmods.dev/badge/skills/zjunlp/mechanist/amortized-shap/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/zjunlp/mechanist/amortized-shap"><img src="https://agentmods.dev/badge/skills/zjunlp/mechanist/amortized-shap.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 67 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 6,517 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. Third-party audits

NVIDIA SkillSpector pass 7 Sept 2026

How audits are shown

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.

Model	Per session	Once invoked
Fable 5.1	$0.00067	$0.06517
Opus 5	$0.00034	$0.03259
Sonnet 5	$0.00013	$0.01303
Haiku 4.5	$0.00007	$0.00652

Measured 9d ago against content hash d5b050066bd1, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

fastshap 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 9d ago.

The scan reads SKILL.md. This mod also ships 2 executable files (scripts/01_tabular_fastshap_demo.py, scripts/02_normalization_and_utils_demo.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

skills/mechanism-skills/SHAP/amortized-shap/SKILL.md · 727 lines

How it starts

The opening of the file, as written. The whole thing — 727 lines — stays where its author put it; the contents beside it link to each section on GitHub.

FastSHAP Skill

When to Use

Activate this skill when:

You need to generate Shapley value explanations for a predictive model's outputs
You want to train an amortized explainer (neural network) that produces explanations in a single forward pass rather than running KernelSHAP separately for each sample
You are working with tabular data (census/adult-style datasets) and want feature attribution explanations
You are working with image data (e.g., CIFAR-10, ImageNet) and need pixel/superpixel-level explanations
You want to train a surrogate model that accepts masked/missing features to support the FastSHAP training process
You need real-time or batch Shapley value estimates with lower computational overhead than KernelSHAP
Keywords: shapley values, SHAP, model explainability, feature importance, amortized inference, KernelSHAP, surrogate model, FastSHAP, local explanations, XAI, interpretability

Quick Reference

Resource	URL
Paper (arXiv)	https://arxiv.org/abs/2107.07436
GitHub Repository	https://github.com/iancovert/fastshap
TensorFlow implementation	https://github.com/neiljethani/fastshap
Census notebook	https://github.com/iancovert/fastshap/blob/main/notebooks/census.ipynb
CIFAR-10 notebook	https://github.com/iancovert/fastshap/blob/main/notebooks/cifar.ipynb
CIFAR-10 single model notebook	https://github.com/iancovert/fastshap/blob/main/notebooks/cifar%20single%20model.ipynb
Blog: Understanding SHAP/SAGE	https://iancovert.com/blog/understanding-shap-sage/

Installation / Setup

Prerequisites

Python 3.7+
PyTorch (install separately per your CUDA version)
A machine learning model to explain (e.g., LightGBM, XGBoost, sklearn, PyTorch CNN)

Install from Source (Recommended)

# Clone the repository
git clone https://github.com/iancovert/fastshap.git
cd fastshap

# Install the package
pip install .

Install Dependencies for Notebooks

Read the full file on GitHub · 727 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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.

9d ago First seen · 727 lines · 67 tokens per session scan A d5b050066bd1

Subscribe to this mod's changes

fastshap is a skill published in the GitHub repository zjunlp/Mechanist (74 stars, last pushed 13d ago), licensed MIT. It adds 67 tokens to every session and 6,517 once invoked, about $0.0003 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 2026-08-30.

Data and AI Mechanistic Interpretability agents ai-scientist ai-systems artificial-intelligence large-language-models

skypilot-multi-cloud-orchestration

Multi-cloud orchestration for ML workloads with automatic cost optimization. Use when you need to run training or batch jobs across multiple clouds, leverage spot instances with auto-recovery, or optimize GPU costs across providers.

OpenRaiser/NanoResearch · 51 tokens

create-atomic-context-provider

Build a BaseDynamicContextProvider that injects a named, titled block into an agent's system prompt at every run() — current time, user identity, retrieved RAG docs, session state, cached DB schema. Use when the user asks to "add a context provider", "inject X into the prompt", "give the agent dynamic context", "wire…

Eigenwise/atomic-agents · 106 tokens

synalinks

Use for anything involving the Synalinks neuro-symbolic LM framework (Keras-inspired) — DataModel/Field/Input, JSON operators (+ & | ^ ), synalinks.ops, LanguageModel/EmbeddingModel and provider prefixes (openai/anthropic/ollama/groq/openrouter/bedrock/...); the Program class and its four building APIs…

SynaLinks/synalinks-skills · 290 tokens

fastshap