Repository navigation
Consolidate overlapping Nextflow skills (41 → 28) - #1
Merged
Merged
Conversation
Make build-nextflow-pipeline the single construction entrypoint so registry composition and source conversion do not compete for activation. Keep the registry-specific approval, module validation and end-to-end checks in a selectively loaded playbook. Replay consolidation in the importer, preserve supporting assets and original source provenance, and redirect callers to the surviving skill. Add CLI tests for asset retention, routing, replay and explicit failures. Validated with consolidation tests, provenance checks, strict Claude skill validation and the structural skill-quality gate for construction.
Remove competing Python, R and notebook entrypoints while retaining language-specific guidance and R reference assets under construction. Use one conversion contract: audit dependencies and hidden state, preserve tested algorithms, split only meaningful tasks and compare scientific outputs with a representative source baseline. Resolve contradictory wrap-versus-decompose instructions and preserve within-task threading. Validated package structure, provenance, replay tests and construction skill structure. This does not claim executed scientific equivalence.
Make nextflow-schema own parameter discovery, schema generation and nf-validation migration instead of exposing three overlapping skills. Retain the four-artifact triage handoff and migration checks for special keywords, help behavior, samplesheet tuple shape and runtime validation. Keep supporting migration documentation and mark its compatibility warning as historical rather than current authority. Validated consolidation tests, provenance, strict Claude skills and the structural skill-quality gate for nextflow-schema.
Use nf-test for both authoring and repairing tests so failure triggers have one entrypoint. Preserve the evidence-based repair protocol as a conditionally loaded playbook instead of duplicating test guidance. Keep regression, intentional change, nondeterminism and environment classification. Require independent expectations and affected-test selection before updating snapshots rather than hiding failures. Validated consolidation tests, provenance, strict Claude skills and the structural skill-quality gate for nf-test.
Remove nf-debug as a competing entrypoint and route lint, config and compilation requests through repair-workflow's static diagnostics path. Keep local and Platform runtime investigation as distinct capabilities. Respect diagnosis-only requests, distinguish profile listing from merged configuration validation, and require runtime evidence beyond lint or preview before claiming a repair is verified. Validated consolidation tests, provenance, strict Claude skills and the structural skill-quality gate for repair-workflow.
Fold nf-run-history into nextflow-history because both investigate the same run records, logs and traces. Keep narrative analysis as an on-demand reference rather than another overlapping activation choice. Preserve time-range and record-completeness checks, evidence-based recurring failure analysis, cache retention safeguards and the distinction between correlation and confirmed cause. Data lineage remains separate. Validated consolidation tests, provenance, strict Claude skills and the structural skill-quality gate for nextflow-history.
Route strict syntax, boolean CLI typing and 25.04 compatibility through migrate-nextflow-code instead of maintaining three additional entrypoints. Preserve syntax examples, release notes and executable legacy scanners under conditionally loaded references with their original provenance. Resolve installed and target releases first and avoid forcing 26.x requirements onto a 25.04 migration. Keep staged baseline, test and output comparison gates instead of treating a parser pass as equivalence. Validated consolidation tests, provenance, strict Claude skills and the structural skill-quality gate for migrate-nextflow-code.
Remove seqera-mcp as a standalone skill and keep a short connection and discovery reference beside launching. Point all callers at that reference without treating it as another callable tool or skill. Delete mirrored API parameter examples: live tool descriptions and returned schemas own operation names, field shapes and authentication. Retain host-managed OAuth, capability-gap reporting, resume identifier safety and post-mutation readback. Debugging and launch skills remain. Validated provenance, strict Claude skills and connection-reference replacement/removal tests in the consolidation suite.
Stop expanding every pipeline step into multiple tool implementations. Use the requested tool or one suitable default; research alternatives only when requested or when the selected tool is blocked. A comparison produces candidates, not implementation authority. Require explicit approval for additional branches and advertise only implemented choices in schema enums. Continue verifying every approved branch. Replay these behavioral edits during import and test their idempotence. Validated consolidation tests, provenance, strict Claude skills and the structural gates for construction and alternative-tool research.
Rebase local Markdown links when skill content moves into nested reference directories. Keep internal asset paths stable while adjusting links to other surviving skills and preserving anchors. This prevents migration references from losing their test-repair and static-diagnostic handoffs after consolidation. Add an independent CLI regression test and regenerate the affected provenance hashes. Validated all seven consolidation tests and source hashes.
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Reduce the bundled skill catalog from 41 to 28 by consolidating overlapping entrypoints while retaining useful guidance as selectively loaded playbooks.
nextflow-schema.nf-testand static diagnostics intorepair-workflow.seqera-mcpwith a short connection reference; live tools own API schemas.Each of the eight consolidation bullets has its own commit and rationale. Separate commits cover opt-in branching and relative-link rebasing.
The importer replays the consolidation plan, retains supporting assets and original provenance, and fails explicitly on upstream wording drift. Scientific-output checks, version-aware migration and diagnosis-only boundaries remain explicit.
Scope
This is the 41 → 28 consolidation, not the proposed 11-skill default pack. Further helper consolidation and optional infrastructure/nf-core/plugin/history packs belong in a follow-up PR. No MCP Skills-extension deployment or Portal ownership migration is implemented here.
Validation
Packaging and routing are validated; no model-based behavioral evaluation or scientific pipeline equivalence is claimed.
Related ownership proposals: seqeralabs/mcp#390 and seqeralabs/portal#1635.