chowkidar 0.6.0

Local-first LLM model deprecation watchdog — scans your configs, alerts on sunsets, and instructs your IDE to update.

Chowkidar

Chowkidar is a secure, local-first LLM model deprecation watchdog. It scans your codebase and configuration files for LLM model references, cross-references them with a locally cached deprecation database, and alerts you before models sunset—safely guiding updates before production breakage.

Everything runs on your machine. Zero data exfiltration.

---

Core Features

• Multi-Format Scanner & Structured Writers: Scans and parses model strings in .env, JSON, YAML, TOML, docker-compose, and source code. At the 1-day threshold, safely auto-updates structured configuration files (with atomic writes, backups, and file locking).
• Notification-First Governance & Per-Reference Audit Log: Alerts via native OS toasts and webhooks (Slack/Discord/generic). Every notification and update attempt is logged in a detailed audit ledger.
• Deployment Signal Detector: Analyzes repo evidence (CI, Docker, Kubernetes, Vercel, Terraform) to flag likely deployed environments, preventing blind or risky local writes.
• Cloud Environment Adapters: Explicit, credential-backed adapters for dry-running, updating, and verifying remote secret/config stores on Vercel, Kubernetes, AWS Secrets/SSM, GCP Secret Manager, and Azure Key Vault.
• Unified Risk & Capability Analysis: Guarantees migrations won't degrade your system by verifying context windows, output tokens, vision, tool usage, JSON mode, streaming, and cost impacts.
• AI-Assistant Rules & MCP Server: Generates zero-config rule instructions (.mdc, CLAUDE.md, etc.) to guide Cursor, Claude Code, Copilot, and Windsurf, alongside an interactive MCP server.

---

Installation & Setup

# Install globally via pipx (recommended) or pip
pipx install chowkidar

# Initialize configuration and local sqlite registry
chowkidar setup --skip-slm

# Fetch deprecation updates from providers (OpenAI, Anthropic, Google, Mistral)
chowkidar sync

# Verify native notifications and setup clickable HTML reports
chowkidar test-notify

---

Command Reference

Command	Description
chowkidar setup	Initialize config (~/.chowkidar/config.toml), registry DB, and SLM
chowkidar sync	Scrape latest model deprecation and sunset schedules from providers
chowkidar scan [PATH]	Find and extract all LLM model references across your codebase
chowkidar check [PATH]	Check found references against the local deprecation database
chowkidar status	Show watched projects, sync freshness, and daemon health
chowkidar watch <PATH>	Register a directory path for periodic background scans
chowkidar unwatch <PATH>	Unregister a directory from background monitoring
chowkidar pin <MODEL>	Suppress notifications for a specific model ID (keeps existing value)
chowkidar unpin <MODEL>	Re-enable notifications for a previously pinned model
chowkidar snooze <MODEL> --days N	Temporarily silence deprecation warnings for a model
chowkidar daemon	Start the periodic background monitoring loop (runs every 4 hours)
chowkidar install-service	Install and register the daemon as an OS-native background service
chowkidar uninstall-service	Remove the OS-native background daemon service
chowkidar logs	Stream or tail daemon execution logs
chowkidar mcp	Start stdio MCP server for rich Cursor, Claude Code, and Copilot queries
chowkidar config	Read or set configuration variables (auto_update, webhook_url, etc.)
chowkidar update	Interactively review and apply recommended model updates to structured configs
chowkidar gate	CI-friendly checker; exits with a non-zero code if critical sunsets are found
chowkidar cost	View pricing comparisons and FinOps savings between models
chowkidar diff <OLD> <NEW>	Perform direct comparison of context size, pricing, and features
chowkidar optimize	Search scanned files for cheaper recommended drop-in model replacements
chowkidar report	Generate comprehensive Markdown, JSON, or interactive HTML reports
chowkidar predict	Estimate model sunset probability based on historical release patterns
chowkidar dashboard	Open a lightweight terminal TUI to visualize model states across projects
chowkidar test-migration	Mock prompt completion outputs on old vs new models before migrating
chowkidar test-notify	Fire a mock native desktop alert and test report click-through action
chowkidar rules write	Generate local prompt files (.mdc, CLAUDE.md) instructing AI agents
chowkidar rules clean	Safely wipe Chowkidar-generated rule instructions from the project
chowkidar slm status	Inspect Ollama connection status and system hardware profile
chowkidar slm choose	Analyze CPU/RAM and configure/pull the best small local model (SLM)
chowkidar slm unload	Unload the SLM from memory to instantly reclaim system RAM
chowkidar hook install	Register a lightweight terminal cd warning hook
chowkidar hook uninstall	Uninstall the terminal cd warning hook

---

Editor Integration

Passive AI Rules (Zero-Config)

AI editors auto-discover instructions in your project workspace. Chowkidar outputs non-destructive rule tables wrapped in <!-- chowkidar --> comments:

• Cursor: .cursor/rules/chowkidar-alerts.mdc
• Claude Code: .claude/rules/chowkidar-alerts.md
• VS Code / Copilot: .github/copilot-instructions.md
• Windsurf: .windsurfrules

MCP Server (Active)

Configure the stdio MCP server in your IDE's configuration file:

{
  "mcpServers": {
    "chowkidar": {
      "command": "chowkidar",
      "args": ["mcp"]
    }
  }
}

---

Security & Local Safety

• Privacy First: No code, project paths, keys, or configurations are ever sent to external APIs.
• Safe Writes: Modifying configuration files requires setting auto_update = true in your config. Every update atomic-writes via a temp file and saves a .chowkidar.bak file for automatic rollback.
• Concurrent-Safe: Uses system-level filelock to protect files from concurrent daemon/CLI writes.

License

MIT