Refresh a repository's shared QStack plan template from the installed qstack-plan-to-html template without rewriting any plan or execution record. Use when plan board styles or scripts are stale, when qstack/compoundengineering/plans/.template/v1 differs from the installed skill, or automatically from…
Rewrite the immediately previous assistant answer for a named reader, such as a product manager, CEO, or CMO, optionally framed around a named product. Use when the user invokes /qstack-explain-for [role] [product], says the last answer is too technical, too low level, or too narrow, or asks for it at a higher level…
Catch the code that got built because it seemed useful rather than because the goal required it. Restate the goal, test each assumption against evidence, then remove, simplify, optimize, or automate in that order and never further than the goal needs. Does not cut validation, error handling that prevents data loss…
Build a clickable static HTML prototype of one or two screens and keep it in the repo under qstack/compoundengineering/prototypes/ /, or attach it to a QStack plan, where it is embedded in plan.html through an iframe with a link to open it in its own tab. Works from a brief in the conversation, a design doc, or an…
Take every open question the previous assistant answer left for the user and ask each one through the host's structured question tool, rephrased for a named reader (product manager by default), with an ELI10 version, lettered options, and one recommendation. Use when invoked as /qstack-ask-as-questions [role], when…
Get a finished branch ready for a human reviewer and then, only on the user's word, open the pull request. Checks the branch is shippable, reads the repository's own landing rules, and writes a handoff the reviewer reads before the diff: reading order, a change map in pseudocode, a diagram only where flow changed…
Run a small task from the conversation as a fan-out of parallel subagents, with the expensive model reserved for orchestration and a cheaper named model doing the work. Takes the orchestrator model and the worker models as arguments, splits the task into independent units, dispatches one subagent per unit on a worker…
Pick the model and reasoning effort for one piece of work, from the models this machine can actually reach and the benchmark scores that exist for them. Discovers the catalogue from the CLI's own proxy configuration, scores it with the Artificial Analysis Data API when a key is present, maps it to strong / standard /…
Rewrite the immediately previous assistant answer in ASD-STE100 Simplified Technical English at a chosen strictness, keeping every fact, the same reader, and roughly the same length. Use when the user invokes /qstack-ste, optionally with a level from 1 to 100 such as /qstack-ste 80, or when the user's whole message is…
Review a pull request, a branch, the working tree, a directory, or the whole repository for exploitable security defects and report findings with a computed score. Carries its own diff review, adapted from Anthropic's claude-code-security-review action, and asks on every run which installed security collections to…
Run a task talked through in the conversation as a lineup of roles on models and efforts chosen by question, through T3 Code's cross-provider delegation, with a reviewer from another model family pressing every unit until it comes back clean. No plan, board, or execution record. Use as /qstack-t3-gegenpress [task]…
Take the finished work in a T3 Code thread all the way to merged: simplify it, open its pull requests, have another model review each one under the /qstack-review contract round after round while this thread fixes what it confirms, and merge each pull request once a round scores 5/5 at its current head. Defaults to…
★not rated 7 3d agoA121 tokens
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: