菜单
Carry an existing Galaxy workflow's tests forward as a regression baseline and augment them for a change-set's deltas, emitting a Galaxy test plan.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Triage failing Galaxy run outputs; classify the failure surface and capture evidence before recommending repairs.
Synthesize a Galaxy workflow test plan from a free-form summary and the Galaxy design briefs.
Assemble Galaxy workflow test fixtures and assertions.
Execute a workflow's tests via Planemo; emit structured pass/fail and outputs.
Convert one nf-core module dir into a Galaxy tool wrapper (tool.xml + macros.xml + _provenance.yml + remote-URL <test> blocks).
Path from a CWL Workflow to a Galaxy gxformat2 workflow — orchestrates the Foundry skills of the CWL → GALAXY pipeline in order, in a per-run working directory.
| name | changeset-to-galaxy-test-plan |
| description | Carry an existing Galaxy workflow's tests forward as a regression baseline and augment them for a change-set's deltas, emitting a Galaxy test plan. |
Follow the procedure below and use the artifact/reference sections as the runtime contract.
summary-galaxy-workflow. Schema: summary-galaxy-workflow. Produced by summarize-galaxy-workflow. Structured summary of the existing workflow from summarize-galaxy-workflow; its tests[] are the regression baseline to carry forward, and its resolved input/output labels are the anchors baseline assertions bind to.galaxy-workflow-changeset. Produced by interview-to-galaxy-workflow-changeset. Reviewed, step-anchored change-set from interview-to-galaxy-workflow-changeset; names which edits change observable behavior, hence which baseline cases to update and which new assertions to add.galaxy-test-plan as galaxy-test-plan.yml. Format: yaml. Schema: galaxy-workflow-test-plan. Reviewable Galaxy workflow test plan (see galaxy-workflow-test-plan): baseline cases carried forward as test-evidence plus change-set-driven cases/assertions, with job inputs, expected outputs, assertion intent, fixture provenance, label status, unresolved mappings, and omissions.references/schemas/galaxy-workflow-test-plan.schema.json: Schema file copied verbatim into the bundle. Output contract: the emitted plan conforms to galaxy-workflow-test-plan. Cast bundles the JSON Schema so the skill carries its output shape; validate with foundry validate-galaxy-workflow-test-plan.references/schemas/summary-galaxy-workflow.schema.json: Schema file copied verbatim into the bundle. Input contract: read the existing workflow's tests[] (the regression baseline) and its resolved input/output labels so baseline assertions bind to real labels.references/notes/galaxy-workflow-testability-design.md: Research note copied verbatim into the bundle. Decide which change-set-exposed outputs and promoted checkpoints make meaningful new assertions, and how much a baseline assertion must loosen when an edit intentionally changes an output. Use when: adding assertions for a newly exposed output or a new step, or reconciling a baseline assertion an edit invalidates.references/notes/iwc-shortcuts-anti-patterns.md: Research note copied verbatim into the bundle. Distinguish accepted IWC-style test shortcuts from assertion smells when loosening a baseline assertion or synthesizing a new one. Use when: considering existence-only, size-only, image-dimension, or tolerant output checks, or recording an omission.references/notes/iwc-test-data-conventions.md: Research note copied verbatim into the bundle. Record job-input fixtures for change-set-added inputs (remote-URL-first locations, hashes, collection shapes) as fixture provenance, reusing the baseline's existing fixtures unchanged. Use when: a change-set adds a workflow input that needs new test data, or when recording provenance for a new fixture.references/notes/planemo-asserts-idioms.md: Research note copied verbatim into the bundle. Pick the assertion family and tolerance for each change-set-driven expected output by output type. Use when: turning a change-set behavioral delta into assertion intent and a tolerance.references/notes/planemo-workflow-test-architecture.md: Research note copied verbatim into the bundle. Keep the plan addressable by stable labels and artifacts Planemo can connect back to invocations, jobs, and outputs. Use when: recording the labels and checkpoints the downstream test must address.references/schemas/tests-format.schema.json: Schema file copied verbatim into the bundle. Use the Galaxy workflow tests schema as the assertion-family vocabulary when carrying forward or synthesizing assertion intent. Use when: naming an assertion family or compare operator for a carried-forward or change-set-driven expected output.galaxy-test-plan.yml before returning it: run foundry validate-galaxy-workflow-test-plan galaxy-test-plan.yml from @galaxy-foundry/foundry. If the command is not on PATH, run npx --package @galaxy-foundry/foundry foundry validate-galaxy-workflow-test-plan galaxy-test-plan.yml. This checks artifact galaxy-test-plan against the galaxy-workflow-test-plan schema.Produce a Galaxy workflow test plan for the update pipeline: carry the existing workflow's tests forward as a regression baseline and augment them for the change-set's behavioral deltas. The output is a reviewable YAML handoff conforming to galaxy-workflow-test-plan, not a concrete tests-format file — implement-galaxy-workflow-test authors the final *-tests.yml from it. This skill is the update pipeline's analogue of nextflow-test-to-galaxy-test-plan / freeform-summary-to-galaxy-test-plan: the dedicated test-plan producer every Galaxy-targeting pipeline places before the implement step.
The starting point is real test evidence — the existing workflow's own tests[], captured in summary-galaxy-workflow. Most of the plan is those cases carried forward, so its dominant basis is test-evidence: set source.kind: galaxy and source.derived_from: test-evidence. source.derived_from is the plan's dominant basis, not a "was anything mixed" flag — the carried-forward baseline dominates, so it reads test-evidence even though the plan also carries synthesized change-set cases. The mix shows through at the finer grains: baseline cases keep test_cases[].derived_from: test-evidence with evidence: test-evidence on their assertions, while change-set-driven cases carry test_cases[].derived_from: intent (or mixed when a carried-forward case gains a change-set-driven assertion) and their assertions carry evidence: intent (raise to test-evidence only where the change-set pinned a concrete expected value). mixed is valid only at the case level, never at source.derived_from. Unlike a freeform-synthesized plan, the baseline's workflow-label bindings are label_status: resolved — the summary read them off the concrete existing workflow, so they are known, not assumed.
Walk the change-set and touch only the assertions its edits reach:
change-parameter, replace-tool, or add-step that alters an existing output's content means the baseline assertion on that output is now wrong. Update it — tighten to the new expected value where the change-set gave one, or loosen it (e.g. presence/format where a content check no longer holds) and record the loosening in omissions[] with a rationale. Never delete a baseline case to make it pass; a regression that silently drops coverage is the failure mode this pipeline exists to avoid.expose-output / add-output needs a new expected_outputs[] entry asserting the promoted output; a new step whose result is observable needs assertion intent for it; a new add-input needs a job_inputs[] binding and a fixture. Bind change-set-added labels at label_status: assumed (the concrete labels settle downstream once the per-step loop resolves the step) and reconcile them in implement-galaxy-workflow-test.rewire, relabel, or remove-step that does not change an observable output changes no assertions; carry the baseline through untouched.Change only the assertions the change-set reaches. Do not add assertions for untouched regions beyond what the baseline already covers, and do not tidy or re-tolerance a baseline assertion no edit touched — the update pipeline's contract is that untouched behavior is verified exactly as before. Gratuitous churn in the plan propagates into gratuitous test churn downstream.
Reuse the baseline's existing fixtures verbatim — they already ship with the workflow and are the regression data. Only a change-set that adds a workflow input needs new test data: record it with iwc-test-data-conventions (remote-URL-first, storage: unresolved with provenance when the input names data only by description), and leave resolution to the harness test-data step / find-test-data and implement-galaxy-workflow-test. When an added input's data cannot be settled here, add an unresolved[] entry (with blocking: true when the new case cannot be authored without it) rather than inventing a location.
Keep the plan addressable by stable labels and artifacts (planemo-workflow-test-architecture) so the downstream test, run, and debug skills can connect assertions back to invocations and outputs.