test-engineering
Skill0 starsChanged 28 days agoMITView on GitHub
Define and govern the repository-wide test system: test strategy, risk-based test design methods, test data and environment policy, flakiness policy, contract/property and non-functional test guidance, test metrics, and independent test-design review. Use when establishing or evolving global testing standards, choosing test levels and design techniques for a risk, defining test data/environment/flakiness policy, or reviewing a task's test design for coverage and determinism before QA execution. Do not use for per-task acceptance testing or verdicts (qa), CI pipeline or gate construction (quality-engineering), semantic plan or code review (review), implementation, or planning.
6 other projects use a version of this.
Adopted by
Most recently taken up by
- juninmd/skillstheir own version, first seen Nov 15, 2026
- Lu1sDV/skillsmdtheir own version, first seen Nov 4, 2026
- asmamest/Nexus-Skillstheir own version, first seen Oct 15, 2026
- nafeeur/MaskShifttheir own version, first seen Sep 12, 2026
- weiweimhy/agent-skillstheir own version, first seen Jul 1, 2026
- yuexueyu/anna-agent-skillstheir own version, first seen Mar 18, 2026
Used with
Projects with it also use code-review.
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.
GoogleSkillcode-review
Adversarial systems, concurrency, memory-model, and architecture code review protocol. Use this skill whenever asked to review code, PRs, diffs, or branches in Filament.
Also in 3 projects that use this
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- juninmd/skills★ 2same name
- Lu1sDV/skillsmd★ 2same name
- nafeeur/MaskShift★ 2same name
- yuexueyu/anna-agent-skills★ 2same name
- asmamest/Nexus-Skills★ 1same name
- weiweimhy/agent-skills★ 0same name
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 repository24 other files it gives its agents.
- AGENTS.mdAGENTS.md
- changeSkill
- convergeSkill
- devSkill
- diagnose-windows-sandbox-aclSkill
- plan-linterSkill
- prd-generatorSkill
- qaSkill
- quality-engineeringSkill
- reviewSkill
- spec-generatorSkill
- task-decompositionSkill
- test-engineeringSkill
- diagnose-windows-sandbox-aclSkill
- changeSkill
- convergeSkill
- devSkill
- plan-linterSkill
- prd-generatorSkill
- qaSkill
- quality-engineeringSkill
- reviewSkill
- spec-generatorSkill
- task-decompositionSkill
The file
We split this skill into the rules it gives an agent
4 rules. 0 are common, 2 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
2
Distinctive
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
- Distinctive: line 70
No test code, no source changes, no CI files, no task-bundle artifacts, and no acceptance verdicts.
- Distinctive: line 74
Assess: Read the current suites, defect history, stack constraints and incident data.
- Very specific: line 77
Formulate the strategy: Write the test-strategy document from assets/templates/test-strategy.template.md: scope, quality attributes, risk …
- Very specific: line 84
Optionally review a test design: Use references/test-design-review.md to audit one task's proposed test design independently.
“Common” is measured: kept by at least 6 in 10 of the 4 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
Test Engineering — Global Test System
The test-engineering role owns the repository-wide test system: how the project decides what to
test, which levels and techniques to use, how test data and environments are governed, how flakiness
is contained, which objective metrics matter, and whether a proposed test design is sound.
This skill is global-only. It is a prompt-only standards-and-review workflow: it never accepts,
verifies, or signs off an individual task, never issues QA_PASSED or REWORK_REQUIRED, never
writes tests or 03-qa-verification.md, never edits source, and never runs or replaces CI gates. It
defers per-task acceptance to qa and CI pipeline/gate construction to quality-engineering.
Use when
- Establishing, revising, or ratifying the repository-wide test strategy and quality attributes.
- Choosing test levels (unit, integration, contract, acceptance, E2E) and design techniques for a stated risk, stack, or feature area.
- Defining shared policy for test data, environments, test doubles, flakiness, or test metrics.
- Reviewing a task's or feature's test design independently before QA execution (design review, not acceptance).
- Onboarding a stack or service into an existing risk-based test system.
- Deciding how test evidence and metrics are defined and handed to QA and CI owners.
Do not use when
- You are executing per-task acceptance testing, recording evidence, or issuing a QA verdict — that
belongs to
qa. - You are building or changing CI pipelines, workflow definitions, gates, or runner environments —
that belongs to
quality-engineering. - You need a semantic plan review (Gate 2) or a change/code review — that belongs to
review. - You are implementing product code or authoring the acceptance test suite — that belongs to
devandqa. - You are planning, slicing, or authoring task contracts — that belongs to the planning skills.
Inputs
- Repository context: supported stacks, critical user journeys, existing suites and their commands, and the current CI topology (read-only).
- Upstream product and architecture artifacts (requirements, decisions, technical spec) and task contracts, when assessing risk or reviewing a design (read-only).
- Defect and incident history, production telemetry summaries, flaky-test reports and test-runtime data.
- A draft test strategy or a task's proposed test matrix / test-design note.
- The adopting project's quality attributes, compliance constraints, and stated thresholds.
- The project's agreed test-strategy location (a role rule may fix a default). The skill ships a strategy template; if no strategy exists, create it from the template instead of assuming a file is present.
Outputs
- A shared, versioned test-strategy document: goals, quality attributes, risk matrix, test-level and technique policy, data/environment policy, flakiness policy, metrics, RACI, and adoption roadmap. Its default location is fixed by the repository role rule; adopters may override that default only through the same rule. When no strategy exists, create it from the bundled template. The skill never depends on a suite-level strategy file being installed; the template is the self-contained source.
- Independent test-design review notes with severity-classified findings and exactly one design
marker:
DESIGN_READYorDESIGN_GAPS. - Reusable standards: technique-selection guidance, review checklists, metric definitions, and a filled strategy template for the stack.
- Handoff recommendations addressed to
qa,quality-engineering,dev, and the planning role. - No test code, no source changes, no CI files, no task-bundle artifacts, and no acceptance verdicts.
Workflow
- Assess. Read the current suites, defect history, stack constraints and incident data. Build or refresh the risk register: likelihood x impact per behavior, component, and interface. Identify quality attributes that are currently unverified or thinly verified.
- Formulate the strategy. Write the test-strategy document from
assets/templates/test-strategy.template.md: scope, quality attributes, risk matrix, level mix, technique policy, data/environment policy, test-double policy, flakiness policy, metrics, RACI, and an adoption roadmap with owners and dates. - Define design standards. Convert the strategy into checkable design rules: which techniques are mandatory for which risk class, what "deterministic" means for this stack, acceptable test doubles, required negative/boundary/state coverage, and the evidence each test must carry.
- Optionally review a test design. Use
references/test-design-review.mdto audit one task's proposed test design independently. Report findings with severity and issue exactly one marker:DESIGN_READYwhen no blocker/major remains,DESIGN_GAPSotherwise. This marker is a design-quality opinion; it is not task acceptance and never a QA verdict. - Hand off and measure. Give QA the design findings and the risk-based priorities; give quality-engineering the metric definitions and tiering expectations; publish the strategy and review notes in the agreed location. Revisit metrics and flake data on a fixed cadence and revise the strategy when evidence changes.
Boundaries
| Concern | Owner | Relationship to this skill |
Per-task acceptance, test execution, evidence, QA_PASSED / REWORK_REQUIRED | qa | Receives design standards and review findings; owns the verdict. This skill never issues it. |
| CI pipelines, workflow definitions, gates, runner environments, automated enforcement | quality-engineering | Receives metric definitions and tiering expectations; owns CI execution and gates. This skill never runs them. |
| Semantic plan review (Gate 2) and change/code review | review | Independent of test design; this skill does not review plans or diffs. |
| Implementation, unit tests, refactoring | dev | Receives design standards; this skill never edits source. |
| Task contracts, slicing, acceptance criteria | planning skills | Read-only here; this skill never authors or rewrites contracts. |
| Repository-wide test strategy, design methods, data/environment and flakiness policy, metrics, test-design review | this skill | Owns the standards and the design-quality marker only. |
Design markers (not a verdict)
DESIGN_READY— the reviewed test design is traceable, risk-appropriate, deterministic, and independently executable; no blocker or major finding remains.DESIGN_GAPS— at least one blocker or major finding is open; list each with evidence and the required action.- Both markers describe test-design quality. They are not
QA_PASSED/REWORK_REQUIRED, not task acceptance, and not a CI gate result.
References
references/test-strategy.md— strategy template and checklist: quality attributes, risk matrix, test levels, technique selection, data/environment and test-double policy, flakiness, automation handoff, metrics, RACI, adoption roadmap.references/test-design-review.md— independent checklist for reviewing a task's test design, severity model andDESIGN_READY/DESIGN_GAPSoutput.references/role-boundaries.md— write scope, read-only scope, invariants, and the honest enforcement model (bundled; no external files needed).assets/templates/test-strategy.template.md— fillable strategy template with sections and placeholders.- Repository rules: read the adopting project's agent rules and role rules when present; they may add requirements but are not a runtime dependency of this skill.
Honest limits
This skill is prompt-driven and cannot enforce standards or isolation. Its outputs are only as good as
the evidence supplied to it, and a design review can miss gaps. Where the repository has CI and branch
protection, that is the authoritative layer for gates and roles. Per-task acceptance remains with
qa; this skill provides the global system and an independent design opinion, never a verdict.
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.

