9k tokens
context cost
the whole folder, loaded on every use
32
files
ships runnable scripts
0
copies elsewhere
how many repositories repackaged it
144
stars on the repo
on the repository, not the skill itself
Install
one command, takes just this skill from the repository
npx skills add https://github.com/naodeng/awesome-qa-skills --skill release-testing-workflow
Copy
What comes with it
35 157 bytes besides the instruction
The instruction itself
7 sections, as written by the author
Release Testing Workflow
中文版: See the corresponding Chinese skill.
When to Use
Need a release-window cadence: T-N planning → specialties → RC → Go/No-Go → post-release watch.
Need release gates and a ship evidence pack, with specialty execution handed to type skills.
Workflow
Read and follow prompts/release-testing-workflow.md (timeline, gates, Go/No-Go, handoffs).
Add release date, scope, freeze rules, candidate build, and known defects that change decisions.
After locating the T window, hand off by skill name per reference.md; invoke specialty type skills by name only.
If input is incomplete, still draft a gate board and mark assumptions—never invent pass results .
Core Constraints
Own release timeline and ship decision; hand specialty reports to performance-testing / security-testing / etc.
Timelines may compress; gate criteria may not be deleted.
Go/No-Go requires evidence; conditional Go must be verifiable.
No relative-path links to other skill files.
Progressive Disclosure
Before producing output, read and follow prompts/release-testing-workflow.md.
For step ↔ handoff mapping: read reference.md.
For stage/specialty deep-dives: invoke the matching type skill; do not expand full specialty reports here.
Templates: output-templates/.
Pre-delivery Checklist
[ ] Followed the main prompt’s output structure
[ ] Includes scope/exclusions, T window, gate board, evidence gaps, next skill
[ ] At decision point: Go / No-Go / conditional with rationale
[ ] Did not invent pass results or unknown defect states
[ ] Assumptions and open questions are marked
Common Pitfalls
Do not delete gates just to compress the calendar.
Do not declare RC done before freeze.
Do not write full specialty long-reports inside this skill.
Do not replace conditional-Go terms with vague “keep watching”.