- Category
- Developer Tools
- Rank
- No. 1031Tools index
- Pricing
- Open Source
- Type
- TOOL
- Builder
- getsentry
- GitHub
- 230 stars
- Latest release
- 3.0.1
- Date
About
Sentry's repository of agent configs and AGENTS.md conventions for using coding agents on real production codebases.
What it does
dotagents turns one declarative setup into working skills, plugins, MCP servers, hooks, and subagents across several coding tools. It installs dependencies globally or per project, records managed state, creates each tool’s expected configuration, and can reconcile, inspect, or repair that state later.
Why it's ranked here
This is a strong fit for teams whose agent setup has spread across multiple clients. Its value comes from explicit scope rules, broad source support, trust controls, lock state, offline reconciliation, and repair commands. The repository also exposes the underlying discovery and validation pieces as a reusable library.
What's good
The design respects configuration ownership. Generated JSON uses sidecars, managed links use reserved markers, and existing client JSON remains unchanged. Project commands cannot silently fall back to global state. Sources may be pinned, local, hosted on common forges, or served over well-known HTTPS. Tests cover scope resolution, worktrees, submodules, discovery compatibility, and workspace linkage.
Tradeoffs
Feature coverage varies by target. Some clients accept MCP servers, hooks, and subagents, while others support only a subset. Pi plugin skills enter a shared global skill area rather than an isolated plugin installation. A source containing any plugin becomes plugin-only during addition, so standalone skills there are not offered. Standard subagent declarations do not unify model, tool, or permission behavior.
How to use it well
Use it when a team wants repeatable agent dependencies across several editors or coding agents. Commit the project declaration, have collaborators install managed state after cloning or pulling, and run reconciliation or health checks when configurations drift. Choose the reusable core when building custom skill consumers. It does not replace each client’s runtime policy, model selection, or permission system.
Technical notes+
The repository is a pnpm workspace with packages/dotagents/ providing the CLI and host library and packages/dotagents-lib/ providing reusable loading, discovery, source resolution, caching, git safety, and trust validation. packages/dotagents/src/cli/main.ts dispatches init, install, add, remove, sync, list, mcp, trust, and doctor commands, with global scope as the default. packages/dotagents/src/scope.ts separates user and project roots, resolves Git worktrees and submodules, and requires agents.toml for established project operations. Tests are colocated under packages/dotagents/src/, while packages/dotagents/vitest.config.ts and packages/dotagents-lib/vitest.config.ts select src/**/*.test.ts in a Node environment.
Observed
- License
- MIT
- Primary language
- TypeScript
- Packaging
- pnpm workspace publishing @sentry/dotagents and @sentry/dotagents-lib
- Install surface
- Runs through npx; requires Node.js 20 or newer
- Interfaces
- Command-line tool, host library, and reusable core library
- Agent targets
- Claude, Cursor, Codex, Grok, VS Code, OpenCode, and Pi
- Testing structure
- Vitest tests are colocated with TypeScript source files
Read from README.md, package.json, packages/dotagents/vitest.config.ts, packages/dotagents-lib/vitest.config.ts, packages/dotagents/src/index.ts, packages/dotagents/src/scope.ts, packages/dotagents-lib/src/index.ts, packages/dotagents/src/index.test.ts, packages/dotagents/src/scope.test.ts, packages/dotagents/src/lib-smoke.test.ts, packages/dotagents/src/cli/main.ts, packages/dotagents/src/cli/index.ts, packages/dotagents/src/utils/index.ts, packages/dotagents/src/agents/index.ts, packages/dotagents/src/config/index.ts.
What it can do
Provide pre-configured agent settings for production codebases
Production codebase repository → Agent configuration files
Define coding agent conventions and standards
Development team requirements → AGENTS.md convention documentation
Enable automated code analysis on real codebases
Source code files → Code analysis results
Configure coding agents for specific project contexts
Project specifications and codebase structure → Customized agent configurations
Standardize agent deployment across production environments
Production environment specifications → Deployment-ready agent configurations
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.