define
Skill0 starsChanged 3 months agoMITView on GitHub
Conversational spec mode — the main session interviews the user, round by round, until the product definition closes; then writes a spec file with verifiable acceptance criteria, ready for /squad:build. Use when the user wants to define, spec, or shape a feature or product, or when a request is too vague to build directly. Also resumes/revises an existing spec file.
11 other projects use a version of this.
Adopted by
Most recently taken up by
- vladcheck/agentictheir own version, first seen Nov 8, 2026
- withpiper/pipekittheir own version, first seen Nov 8, 2026
- vladcheck/skillstheir own version, first seen Nov 1, 2026
- sorawit-w/agent-skillstheir own version, first seen Nov 1, 2026
- marcuskrogh/skillstheir own version, first seen Oct 13, 2026
- Trevoke/org-gtd.eltheir own version, first seen Oct 8, 2026
Used with
Projects with it also use verify.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
Microsoft AzureSkillverify
End-to-end app testing pipeline for browser-based web apps on cloud browsers via Azure Playwright Workspaces: verify the live app, author tests for working flows, heal failing tests, then re-run with video and produce one consolidated report. Three phases (app verification, test authoring, test healing) plus a scoped final sanity run; each phase emits a sub-report. Use when asked to "verify my app", "make sure my app works and has good test coverage", "do a full QA pass", "add tests where I'm missing coverage", "fix failing tests and add new ones", "smoke test before deploy", "QA my deployment", or any combined verify-and-test ask. Requires Azure CLI (`az login`) and a Playwright Workspace resource ID — bootstraps configs on first invocation. Browser-based web apps with Playwright only — not for unit/API tests, native mobile/desktop, load/perf, or full accessibility audits.
Also in 3 projects that use this
Other versions11 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 11
- Trevoke/org-gtd.el★ 485same name
- SensorsIot/Embedded-AI-Harness★ 235same name
- zz41354899/goodux-skills★ 30same name
- saleh-alhaddad/itqan-engineering★ 10same name
- withpiper/pipekit★ 6same name
- andeeplus/aplus★ 2same name
- sorawit-w/agent-skills★ 2same name
- marcuskrogh/skills★ 1same name
- vladcheck/agentic★ 1same name
- vladcheck/skills★ 1same name
1 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository7 other files it gives its agents.
- breaking-changeSkill
- buildSkill
- final-validationSkill
- fixSkill
- refactorSkill
- reviewSkill
- semantic-architectureSkill
The file
We split this skill into the rules it gives an agent
7 rules. 0 are common, 6 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
6
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 29
Read before asking: Check the project's context sources first (CLAUDE.md, README.md, existing specs, overview command if documented) — …
- Distinctive: line 32
Show convergence: After each round, reflect a compact running summary of the spec so far, marking what is still open.
- Very specific: line 34
Respect the pace: If the user pauses or the session ends before closing, save the spec with status: draft so nothing is lost — it can be …
- Distinctive: line 38
When the checklist passes (and only then), render the full spec for a final look, then write it:
- Distinctive: line 49
Never start implementing, and never spawn the delivery agents from here — this skill ends at the artifact.
- Distinctive: line 50
Only the spec file (and its directory) is written.
- Distinctive: line 51
If the user says "just build it" before the spec closes, close early ONLY if the checklist genuinely passes.
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Squad define
Input: $ARGUMENTS — an idea to define, or the path to an existing spec file to resume or revise.
Product definition is a conversation, and the user talks to YOU: run the interview directly in this session — by design, no subagent is involved here. Your single deliverable is a closed spec file — you write no code and change nothing else.
The definition doctrine
Turn the idea into a clear, bounded, executable definition, minimizing implicit decisions. A definition is closed only when all of these exist:
- Problem to solve — what hurts, for whom, in what context
- Affected user(s) and their expectations
- Objective and success signal (even qualitative)
- INCLUDED scope — exactly what is covered
- EXCLUDED scope — explicit, mandatory
- Priority with justification
- Acceptance criteria — numbered, individually verifiable, functional language
- Known risks (technical, user, data, performance)
Hard rules: no ambiguous definitions; if the idea is really several requests, split it into separate specs and ask which goes first; if you cannot write verifiable acceptance criteria, the spec is NOT ready; question anything that adds no clear user or business value.
How to interview
- Read before asking. Check the project's context sources first (
CLAUDE.md,README.md, existing specs, overview command if documented) — never ask what the code or docs already answer, and never propose what already exists. - Small rounds. 2–4 questions per round, the most scope-blocking first. Do not dump the whole checklist at once.
- Propose, don't just ask. For each question, offer the answer you would pick so the user can simply confirm or correct. Fast convergence beats exhaustive surveying.
- Show convergence. After each round, reflect a compact running summary of the spec so far, marking what is still open.
- Challenge. Vague answers get pushed back ("how would the tester verify that?"). Contradictions with earlier answers get flagged, not silently resolved.
- Respect the pace. If the user pauses or the session ends before closing, save the spec with
status: draftso nothing is lost — it can be resumed later by passing its path to this skill.
The spec file
When the checklist passes (and only then), render the full spec for a final look, then write it:
- Location:
specs/<slug>.mdby default; follow the project's existing convention if one exists (docs/specs/, etc.). - Frontmatter:
title,status: ready(draftif unfinished),date(use thedatecommand). - Body: the 8 sections above, plus Notes for UX / Developer / Tester (only what each needs) and Resolved questions (decisions made during the chat, with their rationale — this is the audit trail).
- The spec is written in English regardless of the conversation language.
Close by telling the user: the spec path, its status, and that /squad:build <path> runs the delivery pipeline whenever they are ready.
Rules
- Never start implementing, and never spawn the delivery agents from here — this skill ends at the artifact.
- Only the spec file (and its directory) is written; nothing else is touched.
- If the user says "just build it" before the spec closes, close early ONLY if the checklist genuinely passes; otherwise say exactly which pieces are missing.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

