manual-qa SKILL.md skillA
manual-qa from drunkendealer/claude-code-qa-cycle is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add drunkendealer/claude-code-qa-cycle/manual-qa@git:20261026.35c77afPins git:20261026.35c77af in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- drunkendealer/claude-code-qa-cycle
- Path
skills/manual-qa/SKILL.md- Versions recorded
- 4
- Latest version
- git:20261026.35c77af, committed Oct 26, 2026
- First version recorded
- Oct 4, 2026
- Size
- 10,638 bytes
- Repository stars
- 3
- License
- MIT
Manual-QA pass over recent changes — files concrete bug reports (blocker/major/minor) for UI, navigation, and cross-feature regressions a tester would catch. Auto-detects the platform (Android, iOS, web) from the changed files and audits against that platform's bar; off-platform changes are judged against existing behavior. Builds and tests the branch first, then optionally drives a device or browser to confirm visual bugs. Read-only; output is a fix list, not a test plan. Triggers on…
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Manual QA Simulate a manual tester reviewing a diff and **file the bugs they would file**. Catch UI edge cases, navigation mistakes, and cross-feature regressions that `/code-review` and unit tests miss. **Read-only** — produces a bug list with severities and fixes. Does not modify code. This is the standalone version of `/regression`'s phase 2. It shares that skill's platform checklists rather than keeping its own — run `/regression` when you want the code pass too. Output is a fix list, not a test plan. The user wants to know *what's broken* and *where to fix it*, in the form of a triage-ready bug report. This skill answers: *"if I shipped this right now, what would QA send back?"* Not a replacement for `/code-review` (correctness/security/perf) or instrumented tests. ## Input `$ARGUMENTS` may carry a platform token (`android` · `ios` · `web` · `general`), a scope token, or both, in any order. Anything given explicitly wins — skip detection for it. Scope (mirrors `/code-review`): - empty or `staged` → `git diff --cached` - `last` → `git diff HEAD~1` - `branch` → `git diff main...HEAD` - a file path → that file …
Read the whole file at its exact version.
How to install
mdr add drunkendealer/claude-code-qa-cycle/manual-qa@sha256:a53ac5775a7f399bThe line at the top of this page pins the label git:20261026.35c77af, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_rmsx3hsico5qkm4u)
0 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 4 versions of this file, between Oct 4, 2026 and Oct 26, 2026, about one every 7 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20261026.35c77af latest | 2026-10-26 | 35c77af | 10,638 B | A | view · diff |
| git:20261004.0a0c331 | 2026-10-04 | 0a0c331 | 10,812 B | A | view · diff |
| git:20261004.c19f39a | 2026-10-04 | c19f39a | 10,686 B | A | view · diff |
| git:20261004.74656e9 | 2026-10-04 | 74656e9 | 10,699 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (10638 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
drunkendealer/claude-code-qa-cycle · 3 stars · license MIT · pushed 2026-11-06 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_rmsx3hsico5qkm4u GET https://markdownregistry.com/api/v1/resolve?ref=drunkendealer/claude-code-qa-cycle/manual-qa GET https://markdownregistry.com/api/v1/blob/a53ac5775a7f399b4f7f50e4c895857b27596235329405f0ffc53784f6cecfda
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.
More from drunkendealer/claude-code-qa-cycle
Every file in drunkendealer/claude-code-qa-cycle