npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@mit-sdg/sync-engine-skill

v1.0.0

Published

Agent Skill for scoped sync-engine application design, implementation, and evidence

Readme

@mit-sdg/sync-engine-skill

@mit-sdg/sync-engine-skill helps a coding-agent coordinator design and deliver bounded application changes on @mit-sdg/sync-engine. It compiles concise role assignments, supports delegated or coordinator-simulated work, and keeps a readable record of decisions, prompts, results, identities, and checks.

Use it for applications built on sync-engine, not for changing the framework itself.

Trust boundary

Role access is prompt-guided unless the selected harness enforces it. The CLI validates and records declared paths, but it is not an operating-system sandbox. Run the skill only where that trust model is acceptable.

Bootstrap runs package installation and setup automatically only for an empty application directory. It uses one temporary staging directory outside the application's ancestor workspace for installation, pinned setup, and final dependency reconciliation, then copies the verified project into the application once. Direct Bun installs disable lifecycle scripts and preserve trusted user configuration, environment, and cache. Failed preparation leaves the application empty and retryable. In an existing application it prints commands for explicit review instead of executing project-controlled code.

The selected application root may use directory aliases. Within that root, bootstrap checks target containment and replaces manifest entries without modifying hard-linked external files. Filesystem checks are not a sandbox against a concurrent hostile writer. Publication errors trigger rollback, but abrupt termination or cleanup failure can leave partial files.

Install and start

Install @mit-sdg/[email protected] through the package or Agent Skill mechanism supported by the coordinator. Load its skills/sync-engine/ directory and expose the sync-engine-skill binary. Load one copy so the sync-engine skill name is unambiguous.

From this repository, load packages/skill/skills/sync-engine/.

sync-engine-skill --help

Then open the coordinator in an application root and request an outcome:

Build a durable team task board with projects, assignments, and a small web UI.
Ask me before implementation.

The coordinator creates or resumes .sync-engine/work/<slug>/, follows its immutable review/execution policy, records consequential decisions, selects only useful roles, validates the result, and reports the evidence. grant init creates role-aware capability grants, harness recommend distinguishes a native execution harness from an outer supervisor, completion checks changed paths against the grant, and work finish catches unfinished records before handback.

Adaptable coordination

Independent design and criticism are recommended when behavior changes. Existing accepted design may go directly to implementation. Explicit user direction can skip, combine, replace, or simulate roles.

When a role is delegated, the default first native message is a short instruction to read its generated prompt file. This avoids transporting a large assignment in the first message while keeping the complete prompt auditable. A simulated role uses the same prompt and result artifacts but records coordinator as executor and independent: false without inventing an agent identity. Same-role simulation repairs use compact continuations.

Roles receive inline authoritative context plus explicit read and write directories. The CLI warns when access exceeds a role recommendation. Roles do not inspect node_modules, framework internals, skills, traces, or unrelated project areas unless a higher-priority instruction explicitly changes the boundary.

Inspect a work item

sync-engine-skill work show <slug>

The overview includes the work brief, active decisions, concise activity, runs, execution modes, statuses, and delegated identities. Detailed timestamped artifacts remain available in the work directory.

Support

Stable 1.x follows Semantic Versioning for the documented CLI and workflow contracts. Only the newest stable 1.x release receives fixes. Keep the skill and its required packages at the matching exact versions in skills/sync-engine/release.json. See the support policy and private security reporting process.

Read next

  • SKILL.md — operational entrypoint.
  • workflow.md — coordination procedure and context selection.
  • harnesses.md — file-backed launch and continuation behavior.