qa-ai-assisted
AI-assisted test generation + curation: 3 skills (ai-spec-coverage-mapper, ai-test-generator, model-based-test-graph-author) and 3 agents (ai-test-curator, ai-test-shallow-coverage-critic, mbt-suite-builder).
Install this plugin
/plugin install qa-ai-assisted@testland-qaPart of role bundle: qa-role-ai
qa-ai-assisted
AI-assisted test generation + curation. AI-generated tests from natural-language specs, coverage gap mapping, and two adversarial reviewers for AI-generated tests: one catches hallucinations / weak assertions / redundancy, the other owns the input-domain coverage audit (equivalence partitioning, boundary values, negative paths) that catches shallow happy-path-only suites. The model-based-test graph authoring skill produces structured state-machine input and runs the full graph-to-suite pipeline.
Components
| Type | Name | Description |
|---|---|---|
| Skill | ai-test-generator | Build-an-X workflow that uses an LLM to generate tests from natural-language specs (acceptance criteria, user stories) - outputs tests with confidence scoring per case (LLM's own self-assessment + heuristics: assertion-quality, naming, completeness), batches uncertain cases for human review, integrates with the team's existing test framework. Critical: AI-generated tests are unreliable without curation; pairs with ai-test-curator (the adversarial reviewer). Use when a team has many AC to convert and wants AI-augmentation, not AI-replacement. |
| Skill | ai-spec-coverage-mapper | Build-an-X workflow that uses an LLM to map existing tests to spec sections - given a spec doc + the test suite, the LLM identifies which tests cover which sections, surfaces uncovered sections (gap), and recommends specific tests to add. Output is a coverage matrix per spec ID. Use as a follow-up to ai-test-generator (which generates tests for new ACs) - this maps the existing landscape and finds what's missing. |
| Skill | model-based-test-graph-author | Build-an-X workflow for model-based testing (MBT) - authors a state-machine model of the SUT (states + transitions + guards + actions), validates the model is connected and complete, and runs the full graph-to-suite pipeline: coverage criterion, covering paths, the path-to-acceptance-criteria bridge ai-test-generator consumes, and suite assembly with confidence tiers plus a curation note. Use when a complex stateful flow (checkout, onboarding, multi-step wizard) needs systematic coverage that ad-hoc tests miss. |
| Agent | ai-test-curator | Adversarial reviewer for AI-generated tests - reads the LLM's output and flags hallucinated APIs (functions / classes / imports the LLM invented), weak assertions (.toBeTruthy() style), redundancy with existing tests, missing setup/teardown, and naming patterns the LLM defaults to. Refuses to mark generated tests "ready" if any high-confidence issue remains. Use as the required downstream gate for ai-test-generator - never merge AI-generated tests without this curator's approval. |
| Agent | ai-test-shallow-coverage-critic | Adversarial reviewer that flags tests covering only the happy path - owns the input-domain coverage audit: per public entry point it scores equivalence partitioning (clustering the literal values tests actually pass), boundary value analysis (n/a when no ordered bound is declared), and error/negative-path coverage (negative-assertion ratio), emitting a PASS / SHALLOW / N-A verdict per axis with evidence. Distinct from ai-test-curator (hallucinated APIs, weak assertions) and test-code-critic (vague matchers). Use as the required downstream gate after any AI-assisted test generation, including ai-test-generator, Copilot-suggested tests, and Cursor-authored tests. |
Install
/plugin marketplace add testland/qa
/plugin install qa-ai-assisted@testland-qaSkills
ai-spec-coverage-mapper
Build-an-X workflow that uses an LLM to map existing tests to spec sections - given a spec doc + the test suite, the LLM identifies which tests cover which sections, surfaces uncovered sections (gap), and recommends specific tests to add. Output is a coverage matrix per spec ID. Scope is mapping tests that already exist and naming the gaps, not authoring tests for new acceptance criteria. Use when a spec doc and a test suite both exist but nobody can say which requirements are actually covered - before a release sign-off, an audit, or a decision about where to spend the next round of test effort.
ai-test-generator
Generates tests from natural-language specs (acceptance criteria, user stories, requirements) using an LLM, with confidence scoring per test case (LLM self-assessment plus heuristics: assertion quality, naming, completeness), batching uncertain cases for human review, and integration with the team's existing test framework. Use when the user asks to generate unit tests from acceptance criteria, convert user stories to test cases, automate test creation from requirements, or augment a spec-driven test suite with AI-generated stubs that are then curated before merge.
model-based-test-graph-author
Build-an-X workflow for model-based testing (MBT) per the canonical definition - authors a state-machine model of the SUT (states + transitions + guards + actions), validates the model is connected and complete, and runs the full graph-to-suite pipeline: pick a coverage criterion, generate covering paths, bridge each path into an acceptance-criterion entry ai-test-generator consumes, and assemble the generated tests into a suite file with confidence tiers and a curation note routing to ai-test-curator before merge. Per Wikipedia (en.wikipedia.org/wiki/Model-based_testing): MBT "leverages model-based design for designing and possibly executing tests." Use when a complex stateful flow (checkout, onboarding, multi-step wizard) needs systematic coverage that ad-hoc tests miss, or when the whole model-to-suite pipeline should run in one coordinated pass.
Agents
ai-test-curator
Adversarial reviewer for AI-generated tests - reads the LLM's output and flags hallucinated APIs (functions / classes / imports the LLM invented), weak assertions (`.toBeTruthy()` style), redundancy with existing tests, missing setup/teardown, and naming patterns the LLM defaults to. Refuses to mark generated tests "ready" if any high-confidence issue remains. Use as the required downstream gate for `ai-test-generator` - never merge AI-generated tests without this curator's approval.
ai-test-shallow-coverage-critic
Adversarial reviewer that flags tests covering only the happy path - no boundaries, no error branches, no negative cases. Owns the input-domain coverage audit: per public entry point it scores equivalence partitioning (clustering the literal values tests actually pass), boundary value analysis (recorded n/a when no ordered bound is declared), and error/negative-path coverage (the negative-assertion ratio), emitting a PASS / SHALLOW / N-A verdict per axis with evidence. Distinct from `ai-test-curator` (hallucinated APIs, weak assertions) and `test-code-critic` (vague matchers): it judges whether the test data spans the input space, not assertion specificity. Refuses to clear a test file unless every applicable axis passes per entry point. Use as the required downstream gate after any AI-assisted test generation - `ai-test-generator`, Copilot- or Cursor-authored tests - or when a test file's cases all look alike and the suite needs a defensible answer on whether it exercises more than one equivalence class.