What makes Agentskill notable?

1.1K downloads. Trust evidence available

How should Agentskill be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

agentCLAWHUBUnverified

Agentskill

Let any agent produce code indistinguishable from the existing codebase.

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

1.1k

Updated

Oct 11, 2026

Version

1.4.0

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 1.1K downloads reported by the source. Last updated 10/11/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 11, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 11, 2026
Adoption signal: 1.1K downloadsadoption · observed Oct 11, 2026
Latest release: 1.4.0release · observed May 6, 2026
Handshake status: UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s17f2vd423cnfs44ybpyjk4zv585kxfb:agentskill

Install using `clawhub skill install s17f2vd423cnfs44ybpyjk4zv585kxfb:agentskill` in an isolated environment before connecting it to live workloads.
No published capability contract is available yet, so validate auth and request/response behavior manually.
Review the upstream CLAWHUB listing at https://clawhub.ai/airscripts/agentskill before using production credentials.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-airscripts-agentskill/snapshot"

Documentation

CLAWHUB

160,000 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: agentskill
description: Let any agent produce code indistinguishable from the existing codebase.
---

# SKILL.md — agentskill

> **Operational spec for agentskill.**
> This file governs _when_ to invoke, _what_ to run, and _in what order_.
> For _how_ to generate `AGENTS.md`, read [`SYSTEM.md`](./SYSTEM.md) — it is the behavioral bible.
> These two files are complementary. Neither is sufficient alone.

---

## Purpose

Analyze one or more code repositories. Extract exact coding conventions. Synthesize a precise, forensic `AGENTS.md` that allows any agent to produce code indistinguishable from the existing codebase.

---

## Generation Modes

agentskill supports two generation modes. The mode determines who authors the
final document:

### AI-led generation (skill mode — this file)

The model synthesizes the final `AGENTS.md` itself. CLI analyzer commands are
used **only for evidence gathering** — to extract repository facts that the
model cannot derive reliably from reading source files alone.

**In skill mode, never call `agentskill generate` to produce the final
`AGENTS.md`.** The model is the author. Analyzer output is the raw material,
not the finished product.

### CLI static generation (operator mode)

The user runs `agentskill generate` directly. The packaged runtime emits
markdown automatically. This is appropriate for deterministic direct generation
without an LLM in the loop.

**Use CLI generation only in non-LLM static/operator workflows where the user
explicitly wants tool-generated markdown rather than AI-authored synthesis.**

---

## Rule: AI Authorship

> **The model authors the final document in skill mode.**

- Do not use `agentskill generate` or `python scripts/generate.py` to produce
the final `AGENTS.md` when operating as a skill or in any AI-assisted
workflow.
- Use analyzer commands (`analyze`, `scan`, `measure`, `config`, `git`, `graph`,
`symbols`, `tests`) to gather repository facts.
- The final generated markdown must be synthesized by the AI from analyzer
evidence, direct source file reads, and supporting documentation.
- Treat analyzer outputs as evidence, not as the final authored document.

---

## Trigger Phrases

Invoke this skill when the user says any of the following — or a close paraphrase:

- _"Generate an AGENTS.md"_
- _"Extract my coding style"_
- _"Analyze my repo for conventions"_
- _"Create a style guide from my code"_
- _"Update my AGENTS.md"_
- _"My agent doesn't write code the way I do — fix it"_

Do **not** invoke this skill for general code review, refactoring, or style advice not tied to generating `AGENTS.md`.

---

## File Ecosystem

| File | Role |
| ------------------------ | -------------------------------------------------------------------------------------- |
| `SKILL.md` _(this file)_ | Operational spec: workflow, scripts, fallbacks, uncertainty handling

docs/reference/README.md

# API Reference

This directory documents the packaged `agentskill/` namespace as shipped.

The public CLI surface is the installed `agentskill` command wired through
`agentskill.main:main`. Analyzer implementations live in `agentskill.commands`,
shared orchestration and generation/update helpers live in `agentskill.lib`,
and reusable low-level helpers live in `agentskill.common`.

Reference pages:

- [`cli.md`](./cli.md): packaged CLI entrypoint, subcommands, and dispatch
- [`commands.md`](./commands.md): analyzer command modules and their primary callables
- [`library.md`](./library.md): orchestration, output, update, generation, and reference helpers
- [`common.md`](./common.md): shared registries, filesystem helpers, and repository walking utilities

This reference is intentionally static and release-oriented. It describes the
current packaged layout and contributor extension points rather than every
private helper.

README.md

# agentskill

[![Main](https://github.com/airscripts/agentskill/actions/workflows/main.yml/badge.svg)](https://github.com/airscripts/agentskill/actions/workflows/main.yml)
[![Release](https://github.com/airscripts/agentskill/actions/workflows/release.yml/badge.svg)](https://github.com/airscripts/agentskill/actions/workflows/release.yml)
![ClawHub](https://skill-history.com/badge/airscripts/agentskill.svg)

Analyze a code repository and synthesize an `AGENTS.md` that lets any agent produce code indistinguishable from the existing codebase.

<p align="center">
  <img src="https://raw.githubusercontent.com/airscripts/agentskill/main/assets/agentskill.png" alt="agentskill" width="1280">
</p>

---

## Table of Contents

- [What It Does](#what-it-does)
- [How It Works](#how-it-works)
- [Supported Languages](#supported-languages)
- [Generation Modes](#generation-modes)
- [Install](#install)
- [Development Checks](#development-checks)
- [Usage](#usage)
- [Repository Structure](#repository-structure)
- [Where Code Goes](#where-code-goes)
- [Developer Workflow](#developer-workflow)
- [File Ecosystem](#file-ecosystem)
- [Examples](#examples)
- [API Reference](#api-reference)
- [Contributing](#contributing)
- [Security](#security)
- [Statistics](#statistics)
- [Support](#support)
- [License](#license)

---

## What It Does

agentskill is not a linter and not a style guide generator. It is a forensic extraction tool. It walks a repository, measures every line, reads every config file, and inspects the commit log — then synthesizes a precise behavioral spec for a code-generating agent.

The output is not advice. It is mimicry instructions.

---

## How It Works

Seven analyzers run in parallel. Each extracts one class of signal that an LLM cannot derive reliably from reading source files alone:

| Analyzer  | What it measures                                                    |
| --------- | ------------------------------------------------------------------- |
| `scan`    | Directory tree, file inventory, suggested read order                |
| `measure` | Exact indentation, line length percentiles, blank line distributions |
| `config`  | Formatter, linter, and type-checker detection with config excerpts  |
| `git`     | Commit prefixes, branch naming, merge strategy, signing             |
| `graph`   | Internal import graph, circular dependencies, most-depended modules |
| `symbols` | Symbol name extraction, naming pattern clustering, affix detection  |
| `tests`   | Test-to-source mapping, framework detection, fixture extraction     |

Analyzer output feeds directly into `AGENTS.md` synthesis. The synthesis step follows the behavioral spec in [`SYSTEM.md`](./SYSTEM.md).

> Check our latest technical article for a deeper dive:
> [Turning Repository Knowledge Into Usable Agent Context](https://dev.to/airscript/turning-repository-knowledge-into-usable-agent-context-4pe4).

---

## Supported Languages

agentskill already ships analyzer coverage and repository e

_meta.json

{
  "ownerId": "kn74ef8bfstn5pmnv2b2jkc73d85jqdn",
  "slug": "agentskill",
  "version": "1.4.0",
  "publishedAt": 1778109397554
}

references/GOTCHAS.md

# GOTCHAS.md — Extraction and Synthesis Errors

> Read this file in full before drafting any section of `AGENTS.md`.
> Every entry here is a failure mode discovered from an actual run.
> When you discover a new one, add it.

---

## Extraction Errors

These errors occur during data collection — the signal is wrong before synthesis even begins.

---

### Keyword pollution

**What happens:** Language keywords (`self`, `cls`, `if`, `for`, `return`) are counted alongside identifier names, inflating snake_case totals and polluting naming pattern analysis.

**Fix:** Filter keywords before classifying names. Do not count anything that appears in the language's reserved word list.

---

### Single-word name ambiguity

**What happens:** A name like `foo` or `data` matches both `camelCase` and `snake_case` classifiers because it has no case transitions. These names dominate short codebases and produce false confidence.

**Fix:** Require at least one case transition or underscore before classifying a name. Single-word names are `other`, not evidence of any convention.

---

### Generated file skew

**What happens:** Vendored files, lockfiles, and generated code have zero comments, uniform indentation, and no meaningful names. Including them distorts every measurement.

**Fix:** Exclude all directories in `SKIP_DIRS` — including `node_modules`, `vendor`, `dist`, `build`, `.eggs`, `site-packages`, and `__pycache__`. Do not include `.lock` files in line length or whitespace analysis.

---

### Test file bias

**What happens:** Test files use different idioms than source files — more `assert` statements, more fixture variables, more repetitive naming. Mixing them into source analysis contaminates naming and error handling measurements.

**Fix:** Analyze test files and source files separately. Only report source-file patterns as codebase conventions. Call out test-specific patterns explicitly under Section 12.

---

### Blank line measurement at file boundaries

**What happens:** The first top-level definition in a file has no predecessor, so the blank line count before it is always zero. Including this in the distribution pulls the mode toward zero even when the real convention is two blank lines between definitions.

**Fix:** Exclude the first definition in each file from the blank-line-between-definitions measurement. Only measure gaps _between_ two definitions, never before the first one.

---

### Import misclassification

**What happens:** stdlib module names that overlap with third-party package names (`email`, `ast`, `typing`) get classified as third-party, and vice versa. This produces incorrect import ordering rules.

**Fix:** Maintain an explicit stdlib module list. Check against it before classifying an import. When uncertain, check the module's origin via `sys.stdlib_module_names` (Python 3.10+) rather than guessing.

---

### Branch inflation from remote tracking refs

**What happens:** `git branch -a` returns both local branches and remote trackin

Github ReposUpdated 2d agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.