shinytoyrobots

Agentic Delivery Suites — AI skill for Claude Code

AI community

Two experimental Claude Code skill suites for building software with AI agents — one accelerates the Scrum sprint, the other asks whether to keep it at all.

How to install Agentic Delivery Suites

This entry records only its repository, not the path inside it, so there is no exact command to give. Open shinytoyrobots/agentic-delivery-suites and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Agentic Delivery Suites does

Two experimental Claude Code skill suites for building software with AI agents — one accelerates the Scrum sprint, the other asks whether to keep it at all.

Alternatives in AI

Paca — AI-native, free, open-source alternative to Jira, Trello, ClickUp & Monday 1.8k ★
Loushang — AI-native agent harness for coding workflows by python: multi-model LLM orchestration, stateful sessions, tool 1.3k ★
SoloDawn — Catches and repairs AI hallucinations before delivery: three quality gates, 31 built-in rules, a self-healing 312 ★

README

Agentic Delivery Suites

Most tools for building software with AI keep the process we already have — the Scrum sprint, with its stories, gates, and fixed roles — and accelerate it. These two [Claude Code](https://claude.com/claude-code) skill suites take opposite answers to a sharper question: once the people in the loop are agents, is the sprint worth keeping at all?

Suite	Premise	In one line
`delivery-team`	AI acceleration of a human sprint	13 sub-agents acting as a complete product-delivery team, orchestrated through an 11-skill sprint lifecycle.
`flow`	AI-first from the ground up	The spec is the source of truth; code is regenerated output. One wide probe generation of parallel variants, multi-objective Pareto evaluation, an evidence-gated ship instead of a calendar — with weight-classed cost envelopes and mandatory token budgets.

`delivery-team` keeps the shape of a Scrum sprint — stories, gates, retros, fixed roles — and makes every seat faster. `flow` discards that scaffolding and rebuilds the pipeline around what LLM agents are uniquely good at (parallel reading, structured scoring, deterministic projection) and bad at (parallel writes, free-form coordination, drifting prose).

**[delivery-team-vs-flow.md](./delivery-team-vs-flow.md)** is the full side-by-side comparison and the best place to start if you're deciding which to use.

What's here

flow/                 # AI-first delivery suite
  commands/           #   slash-command skills (flow-init, flow-spec, flow-generate, …)
  agents/             #   function-shaped sub-agents (orchestrator, evaluator, generator, …)
  context/            #   shared reference docs the skills load
  README.md           #   suite overview
  USAGE.md            #   practitioner's guide
delivery-team/        # Sprint-accelerator delivery suite
  commands/           #   sprint lifecycle skills (dt-start, dt-run, dt-clo