Back to Skills

cli-anything-adguardhome

Command-line interface for AdGuard Home - Network-wide ad blocking and DNS management via AdGuard Home REST API. Designed for AI agents and power users who need to manage filtering, DNS rewrites, clients, DHCP, and query logs without a GUI.

45,315stars4,234forksUpdated 7/14/2026

Security Assessment

Safe(100/100)
Security Score100/100

About cli-anything-adguardhome

CLI-Anything for Codex adapts the full CLI-Anything methodology so that Codex can build, refine, test, validate, and list CLI harnesses that wrap GUI applications or source repositories. The core problem it solves is turning software that normally requires a graphical interface (or a raw source checkout) into a scriptable, agent-friendly command-line tool, while keeping the generated Python harness format identical to the canonical CLI-Anything output. Before doing any work it directs the agent to read the authoritative sources first: `references/HARNESS.md`, the matching mode spec under `references/commands/`, and any applicable `references/guides/`, falling back to a repository checkout or a clone of `https://github.com/HKUDS/CLI-Anything` when installed resources are unavailable.

The skill defines five modes, each backed by its own command spec: Build (`references/commands/cli-anything.md`) performs source analysis, architecture design, implementation, test planning and implementation, documentation, CLI-specific `SKILL.md` generation, and local installation; Refine inventories existing commands and adds high-impact capabilities without removing them; Test runs harness tests against the real backend and updates `TEST.md` only with passing results; Validate checks the harness against a full directory, implementation, test, documentation, packaging, and code-quality checklist; and List discovers installed or generated tools with human-readable or JSON output. A condensed fallback specifies the harness directory structure, required behaviors (real backend preference, Click subcommands plus a default REPL, `--json` output, undo/redo session state, `--dry-run`, locked session writes, the unified `ReplSkin`), testing rules (`test_core.py`, `test_full_e2e.py`, `CLI_ANYTHING_FORCE_INSTALLED=1`), and packaging via `find_namespace_packages(include=["cli_anything.*"])` with `console_scripts` exposing `cli-anything-<software>`.

It targets developers and Codex-driven agents who want reproducible CLI harnesses for GUI apps or repos. Inputs may be a local path or a GitHub URL. A stated requirement is that the full methodology sources should be consulted first, and the condensed rules used only when those sources cannot be retrieved.

FAQ

What inputs does this skill accept?

It accepts either a local source path (such as ./gimp or /path/to/software) or a GitHub repository URL. When cloning is needed, the software name is derived from the local directory name.

What modes are available?

Five modes: Build, Refine, Test, Validate, and List. Each has its own specification file under references/commands/ (cli-anything.md, refine.md, test.md, validate.md, and list.md).

Which sources should be read before implementation?

Read references/HARNESS.md first, then the matching mode spec under references/commands/, and referenced files under references/guides/ only when they apply. If installed resources are missing, fall back to a CLI-Anything repository checkout or clone https://github.com/HKUDS/CLI-Anything.

How is the generated harness packaged?

Using find_namespace_packages(include=["cli_anything.*"]), keeping cli_anything/ as a namespace package without a top-level __init__.py, exposing cli-anything-<software> through console_scripts, and including the packaged CLI-specific skills/SKILL.md.

How should testing be handled?

Write TEST.md before test code, keep test_core.py for unit coverage and test_full_e2e.py for real-file and real-backend workflows, verify rendered output programmatically rather than only exit codes, test the installed command via _resolve_cli(), and run release validation with CLI_ANYTHING_FORCE_INSTALLED=1.

All Files

5 files
agents/openai.yaml0.3 KB
View
scripts/install.sh2.1 KB
View
tests/test_install.sh3.7 KB
View
scripts/install.ps13.1 KB
View
SKILL.md5.9 KB
View

Install cli-anything-adguardhome

Download and extract the skill files to your .claude/skills/ directory.

Quick Setup:

  1. Copy the skill folder to .claude/skills/
  2. Claude will automatically detect and use the skill