Ultraqa
- 55 installs
- 40 repo stars
- Updated August 4, 2026
- akillness/oh-my-skills
ultraqa is a Claude Code/Codex skill that runs an autonomous QA loop (test, diagnose, fix, repeat) via $ultraqa until tests, build, lint, or typecheck gates pass.
About
ultraqa runs an autonomous QA cycling workflow that tests, diagnoses, fixes, and repeats until the quality goal is met or a concrete blocker is proven. A developer invokes $ultraqa to make tests, build, lint, typecheck, or review gates pass using repo-native commands. It matters because it preserves failure evidence and applies the smallest fix each cycle instead of blindly retrying.
- Runs an autonomous test, diagnose, fix, repeat QA loop until the quality goal is met
- Maps a request to one goal: tests, build, lint, typecheck, review, or custom
- Caps at five cycles and stops after three identical failures without new evidence
Ultraqa by the numbers
- 55 all-time installs (skills.sh)
- Ranked #1,195 of 2,153 Testing & QA skills by installs in the Skillselion catalog
- Data as of Aug 5, 2026 (Skillselion catalog sync)
ultraqa capabilities & compatibility
- Capabilities
- test repair · build fix · lint fix · typecheck loop
- Use cases
- testing · debugging · code review
What ultraqa says it does
Run an autonomous QA cycling workflow: test, diagnose, fix, and repeat until the quality goal is met or a concrete blocker is proven.
Stop early if the same failure repeats three times without new evidence.
npx skills add https://github.com/akillness/oh-my-skills --skill ultraqaAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 55 |
|---|---|
| repo stars | ★ 40 |
| Last updated | August 4, 2026 |
| Repository | akillness/oh-my-skills ↗ |
What it does
Run an autonomous diagnose-fix-repeat loop until tests, build, lint, or typecheck gates pass or a blocker is proven.
Who is it for?
Making tests, build, lint, typecheck, or review gates pass via a diagnose-fix-repeat loop
Skip if: Independent implementation lanes (ultrawork), coordinated workers (team), or review-only findings
When should I use this skill?
The user invokes $ultraqa or wants a test/build/lint repair loop
What you get
A passing verification command, or a proven concrete blocker, within the cycle cap
- Passing verification command
- Preserved failure evidence
By the numbers
- up to 5 cycles
- stop after 3 identical failures
- 6 goal types (tests, build, lint, typecheck, review, custom)
Files
UltraQA
Use this exact-name skill for $ultraqa. If the user types $ultaqa, treat it as a typo for $ultraqa and continue with the same QA contract.
When to use this skill
- The user explicitly invokes
$ultraqa,$ultaqa, orultraqa. - The task is to make tests, build, lint, typecheck, or review gates pass.
- The user wants a diagnose-fix-repeat quality loop.
Instructions
Parse one QA goal, run the relevant verification, preserve failure evidence, apply the smallest fix, and repeat until pass or blocker.
Goal Parsing
Map the request to one primary QA goal:
tests: test suite passesbuild: build exits successfullylint: lint exits successfullytypecheck: type checker exits successfullyreview: parallel review of correctness, security, performance, or UI concernscustom: user-specified success condition
Cycle
Repeat up to five cycles:
1. Run the relevant verification command. 2. If it passes, stop and report the passing command. 3. If it fails, preserve the failure evidence. 4. Diagnose the likely root cause. 5. Apply the smallest fix. 6. Re-run verification.
Stop early if the same failure repeats three times without new evidence.
Command Discovery
Prefer repo-native commands from package.json, Makefile, project docs, CI config, or existing scripts. Do not invent a test command when the repo gives a clear one.
Route-Outs
- Use
ultraworkwhen the main job is independent implementation lanes. - Use
teamfor coordinated runtime workers. - Use
code-reviewfor review-only findings without an edit loop.
Examples
$ultraqa --tests$ultaqa make lint and typecheck pass
Best practices
- Use repo-native verification commands.
- Stop after three identical failures without new evidence.
- Report the final passing command exactly.
References
testing-strategiesfor policy-level gate design.code-reviewfor review-only passes.
{
"skill": "ultraqa",
"version": "1.0.0",
"cases": [
{
"prompt": "$ultraqa --tests",
"expected_output": "The ultraqa skill activates, discovers repo-native test commands, runs the QA cycle, and repeats fixes until pass or concrete blocker.",
"checks": [
"Treats `$ultraqa` as a direct trigger",
"Preserves failure evidence",
"Stops only on pass, blocker, or repeated identical failure"
]
},
{
"prompt": "$ultaqa make lint and typecheck pass",
"expected_output": "The typo `$ultaqa` is treated as `$ultraqa` and the same QA cycling contract is applied.",
"checks": [
"Handles the common `$ultaqa` typo",
"Runs lint/typecheck-oriented QA cycling"
]
}
]
}
N:ultraqa
D:Exact-name `$ultraqa` front door for QA cycling: run tests/build/lint/typecheck or review, diagnose failures, fix, and repeat until the goal passes or a concrete blocker is proven. Treat `$ultaqa` as a common typo for `$ultraqa`.
T:Read|Write|Bash|Grep|Glob
G:ultraqa|$ultraqa|ultaqa|$ultaqa|qa|tests|build|lint|typecheck|review|omx|codex
F:Codex CLI|Claude Code|Gemini CLI|OpenCode
S:
1:Map the request to tests, build, lint, typecheck, review, or custom goal
2:Run repo-native verification commands and preserve failure evidence
3:Diagnose and apply the smallest fix
4:Repeat up to five cycles, stopping after three repeated identical failures
R:
ultrawork:Use for independent implementation lanes
team:Use for coordinated runtime workers
code-review:Use for review-only findings without an edit loop
Related skills
TddFollow test-driven development with a strict red-green-refactor loop when creating reliable features or fixing bugs.544k196k
Test Driven DevelopmentEnforce writing failing tests before any production implementation code.181k264kQaRun conversational QA sessions that turn user-reported bugs into well-written, domain-aware GitHub issues without manual ticket writing.179k196kMigrate To ShoehornAutomatically update TypeScript test files that rely on unsafe `as` type assertions by replacing them with type-safe partial objects from @total-typescript/shoehorn.164k196k
Webapp TestingVerify frontend behavior, debug UI issues, capture screenshots, and inspect logs of a running local web application using Playwright.123k165k
Playwright CliRun browser automation, generate element snapshots, inspect DOM attributes, and execute Playwright tests from the terminal.102k12.3kFAQ
How many cycles does it run?
Up to five cycles, stopping early if the same failure repeats three times without new evidence.
What commands does it use?
It prefers repo-native commands from package.json, Makefile, project docs, CI config, or existing scripts.