superlite
A fast software factory workflow for Claude Code: one approval gate, parallel waves, routed models.
A fast software factory workflow for Claude Code: one approval gate, parallel waves, routed models.
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
Use when you have an approved design or requirements for a multi-step task and need an implementation plan, before touching code.
Use in feature mode after the feature plan is approved, to cut the smallest runnable container of an existing app for the…
Use when the user asks to build a feature or component, add functionality, or change behavior, before any implementation code is…
Use in MVP mode (/superlite:mvp), after the .mvp.md feature plan is written, to make the one-file clickable prototype, check it,…
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs
Use when the user wants a topic researched with sources, a sourced report at a chosen language level, or conflicting evidence…
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work
Use when executing a superlite plan that has a %% taskgraph. Runs a ready-queue scheduler over an idle agent pool, with scripts…
Use when implementing any feature or bugfix, before writing implementation code
Use when starting any conversation - explains superlite, how to size a request, and which skill fits which task
Use when agent_specs is on (or the user asks for a spec) to write what done looks like and why before a plan or implementation.
Fresh one-turn final reviewer. Reviews from .superlite/review-packet.md alone (Read only) and returns a verdict plus tiered…
Fix-only agent, tier 3. Diagnoses unknown-cause findings (reproduce, then fix) with a regression test, then records a checkpoint.
Fix-only agent, tier 1. Fixes known-cause findings confined to one module with a regression test, then records a checkpoint.
Fix-only agent, tier 2. Fixes known-cause findings that span modules or a seam, or known probe classes, with regression tests,…
Implements one task from a plan. Use for each task; primed with the shared prime first (READY), then sent its task by message.…
Implements one Kind: mechanical task (fully specified, small, fast tests). Use only for mechanical tasks; primed with the mini…
Reviewer-fixer for one checkpointed task (selective) or one package (batched). Reviews the given diffs against the spec slice and…
Gathers evidence for one research sub-question and records it in the research ledger. Never writes the report.
Switch this project to the AI orchestrator: superlite plans, builds and decides on its own.
Approve the review plan so superlite can lint and build it.
Find the root cause of a bug first, then fix it with a regression test.
Add or change a feature in an existing codebase. Research, a plan you validate, a prototype you try, then the integration build.
Finish the branch: verify first, then offer merge or PR. Never pushes without you.
Show the detailed superlite help: modes, who decides, how a build runs, every command, models, rules, files and commits.
Switch this project to the human orchestrator: you approve a review plan before any build.
Show the superlite intro screen: the logo, how a build runs, and every command.
Plan the smallest working version, assess it on a clickable HTML prototype, then build it with the orchestration flow.
List stored build reports, show one build's report summary, or the history across runs.
Research a topic with parallel agents, a sourced ledger and a report at your level.
Resume an interrupted superlite build from the last checkpointed, tested wave.
Is this you? Claim your profilePut a plugin at the top of Plugins, from $9