testing-task
Skill0 starsChanged 3 months agoMITView on GitHub
Generate the concise Arabic testing-team task (title + description) from the approved implemented behavior, warn separately when deployment was skipped, and close the workflow at completed.
In this repository8 other files it gives its agents.
- codex-reviewSkill
- deploymentSkill
- feature-changelogSkill
- frappe-taskSkill
- git-finalizationSkill
- project-contextSkill
- task-implementationSkill
- task-planningSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 32
No file: Never create docs/ai-context/testing-task-ar.md, .claude/testing-task-ar.md, or any replacement testing-task file anywhere.
- Very specific: line 89
When the stage is deployment_skipped, show this separately, in English, after the preview and outside the Arabic content:
- Distinctive: line 97
Never embed the warning inside the Arabic description.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Testing Task Skill
Produces the final hand-off to the (Arabic-speaking) testing team and closes the workflow.
When to Use
- The
testingaction — only from stagedeployedordeployment_skipped.
Inputs
- The approved implemented behavior:
docs/ai-context/TASK_PLAN.md(Required Behavior + Acceptance Criteria as approved), the feature-changelog entry updated at finalization, the implementation summary. - Deployment status from state.
Outputs
- Arabic
Title+Descriptioncopied to the host clipboard in their original logical Unicode order, ready to paste into the testing team's task-management system, plus a terminal-only visual preview of the same text so it is readable in this terminal — content rules in references/arabic-testing-task-rules.md. - No file. Never create
docs/ai-context/testing-task-ar.md,.claude/testing-task-ar.md, or any replacement testing-task file anywhere. - State:
testing_task = {status: "generated", generated_at};state transition completed. The generated Arabic text itself is never stored in state.
Preconditions
- Stage is
deployedordeployment_skipped; commit verified.
Order of Operations
The clipboard copy comes first, and everything else depends on it:
CLI clipboard copy --previewwith the finished Arabic text on stdin.- Exit code
0→ the command has copied the logical text and printed a visual preview of it. Repeat that preview output verbatim, record the state, transition tocompleted. - Exit code
8(no clipboard) → stop. Print nothing of the Arabic text, write no file, record no state, transition nothing. Show the command's own failure output — it names every method that was checked — and tell the user to install a clipboard utility (wl-clipboard,xclip, orxsel) or run from a session that has one, then ask fortestingagain. Never install a package.
The clipboard is the delivery channel; the preview is only there to be read. This terminal draws text strictly left to right, so the logical Arabic appears backwards in it — the preview is the same text reordered for display, and it is exactly what must not be pasted anywhere. Never print the logical Arabic text yourself, and never copy the preview into the clipboard, a file, or the workflow state. A failed copy therefore blocks the workflow instead of falling back to printing.
Stopping Conditions
- Task copied, previewed, state recorded, workflow
completed→ report the closing summary (commit hash, deployment outcome, and that the Arabic testing task is on the clipboard and previewed above). Done. - Copy succeeded but the command reported that the preview could not be formatted → the hand-off is still complete: pass that English warning on, never print the logical Arabic instead, and continue with the state record and the transition.
- Clipboard unavailable → stage unchanged, nothing recorded, the blocker reported to the user. Not a completion.
Repeating After Completion
The text is kept nowhere, so a user who lost it may ask for it again while
the stage is already completed. Regenerate it from the same approved
behavior, copy it to the clipboard and show the preview again, and change
nothing: no transition (there is none from completed), no state write, no
file. A failed copy at this point changes nothing either — the stage stays
completed.
Deployment-Skipped Warning
When the stage is deployment_skipped, show this separately, in
English, after the preview and outside the Arabic content:
Deployment was skipped. Publish the testing task only after the changes
are available in the testing environment.Never embed the warning inside the Arabic description.
Prohibited
- Writing the title or description to any file, in the application repo or
anywhere else. Older plugin versions saved
.claude/testing-task-ar.md/docs/ai-context/testing-task-ar.md; such a file may still exist in an old application — leave it untouched, never read it, never update it. - Printing the Arabic text, recording state, or transitioning to
completedafter a failed clipboard copy. - Printing the logical Arabic text at any point, or putting the visual preview on the clipboard, in a file, or in the workflow state.
- Reordering, reshaping, or reversing the text yourself: the visual form
comes from
CLI clipboard copy --previewand from nowhere else. - Installing a clipboard package, or working around a missing one with a file, a here-doc into an editor, or any other on-disk detour.
- Mentioning source files, Python functions, or internal implementation in the Arabic text.
- Copying the original task title as the description.
- Describing behavior that was not actually approved and implemented.
- Generating from a stage other than
deployed/deployment_skipped.
Shared Rules
shared-workflow-rules.md.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.

