Distribution Engineer is a skill published by agensi that handles the engineering work involved in taking software from a working local build to a production-grade release. It covers four main areas drawn from the facts.

Channel Strategy: the skill selects the appropriate distribution method — npm, pipx, standalone binaries, or containers — based on the intended audience and runtime environment.

Artifact Engineering: it defines a release contract that codifies versioning logic, install paths, and configuration boundaries, giving the release a clear and repeatable structure.

Quality Assurance: the skill produces verification plans that cover fresh installs, upgrade paths, and smoke tests, specifically targeting the clean-path experience that catches PATH issues and environment assumptions before end users encounter them.

Marketplace Readiness: it prepares manifests, release notes, and metadata required to list skills or tools in AI agent marketplaces.

The output of a session is a Distribution Plan document containing an Artifact Contract, Install UX details with specific CLI commands, Verification Results from clean-path testing, and Rollback Notes. This skill is intended for developers who need structured, stage-separated release workflows — build, package, and publish — rather than generic advice. It does not execute deployments itself; it produces plans and documentation an engineer then acts on.

Use cases

Define a versioning and install-path contract for a CLI tool before its first public release

Choose between npm, pipx, a standalone binary, or a container for distributing an AI agent

Generate a verification plan covering fresh-install and upgrade smoke tests

Prepare marketplace manifests and release notes for listing a skill in an AI agent marketplace

Produce rollback notes as part of a release plan for a library update

Structure a release workflow that separates build, package, and publish stages

When to use it

Shipping an AI agent, CLI, or library and need a structured distribution plan

Deciding which package channel (npm, pipx, binary, container) fits a specific target audience

Preparing a skill or tool for submission to an AI agent marketplace

Need clean-path install verification and rollback documentation before a release

When not to use it

Requires a skill that executes deployments or publishes packages directly — this skill produces plans, not actions

Looking for runtime monitoring or post-release observability tooling

Need database-specific distribution workflows outside software packaging

Project has no packaging or release step and only needs code generation