Nightward — AI skill for Claude Code
Local-first TUI for auditing AI agent state, MCP config, and dotfiles backup safety.
How to install Nightward
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open JSONbored/nightward and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Nightward does
Local-first TUI for auditing AI agent state, MCP config, and dotfiles backup safety.
Alternatives in AI
- Anthropic SDK TypeScript — Access to Anthropic's safety-first language model APIs in TypeScript 2.1k ★
- Open Science — Open Science Desktop — local-first, model-agnostic AI research workbench for macOS, Windows & Linux 1.5k ★
- UltraCode Shim — Give Claude Code's ultracode mode to ANY model you already pay for 422 ★
README
Nightward
[](https://github.com/JSONbored/nightward/actions/workflows/ci.yml) [](https://github.com/JSONbored/nightward/actions/workflows/nightward-policy.yml) [](https://github.com/JSONbored/nightward/actions/workflows/scorecard.yml) [](https://www.bestpractices.dev/projects/12713) [](LICENSE)
Nightward finds AI-tool risks before you sync: MCP risk, local-only state, secret exposure, and reviewable fix plans, all locally.
It scans common Codex, Claude, Cursor, Windsurf, VS Code, Raycast, JetBrains, Zed, Continue, Cline/Roo, Aider, OpenCode, Goose, LM Studio, Ollama/Open WebUI, Neovim, MCP config locations, and repo/workspace AI config; classifies what is portable versus local-only or secret; highlights MCP security findings; and produces redacted analysis signals, fix plans, SARIF policy output, snapshot plans, and dry-run backup plans.
Public docs and the fixture TUI walkthrough live at .
Nightward is read-only by default, but it can run explicit, confirmation-gated local actions such as provider installation, provider enable/disable, scheduled scan install/remove, portable config snapshots, policy ignores, and Nightward-owned report/cache cleanup.
[!IMPORTANT] Nightward is local-first by design: no telemetry, no default network calls, and no cloud dashboard. Write-capable actions are beta operator tools. Users must review previews, confirmations, backups, package-manager behavior, and third-party provider behavior before applying changes. Nightward is provided without warranty, and maintainer
Related Skills
Prompt Storage
pst — local-first personal prompt library CLI. SQLite+FTS5, instant retrieval (pst code-review), variable temp
ClaudeR
Connect RStudio to Claude Code, Codex, Gemini, and other LLM agents via MCP. Multi-agent orchestration, automa
Nexpath
Local-first AI coding workflow for vibe coders, indie hackers, technical founders and product managers — catch
Agf
Local-first TUI to find, search, and resume AI coding-agent sessions across Claude Code, Codex, Grok Build, Ki
Vaultpilot MCP
Safety first. Hardware-verified DeFi for AI agents — designed for when the AI can be compromised.
Agent Settings Backup Script
Git-versioned backup tool for AI coding agent config folders (Claude Code, Cursor, Codex, etc.) with size-base
Related Agents
Light Security ClaudeSettingsAudit
Audit Claude Code config files (settings.json, settings.local.json, .mcp.json) for unsafe or malicious content
My Security Reviewer
Fresh-context security reviewer for agent tooling - skills, subagents, commands, hooks, shell/sync scripts, do
Timps Backup Sentinel
Audit backup health: Time Machine last backup time, all local git repos with uncommitted/unpushed changes, lar