ai4curation/ai-gene-review / .claude/skills/module-curation
module-curation
VerifiedCurate, review, or repair ai-gene-review ModuleReview YAML documents under modules/, including pathway/module boundary setting, parts and variant modeling, annoton-level molecular functions, representative UniProt/PTN grounding, module deep-research provenance, validation, rendering, and project batch updates.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.8 KB |
| agents/openai.yaml | config | 205 B |
| references/modeling.md | markdown | 4.9 KB |
| references/research-and-grounding.md | markdown | 2.4 KB |
| references/validation-and-pr.md | markdown | 2.7 KB |
references/validation-and-pr.md
View on GitHub# Validation And PR Checklist
## Required Checks
Run these for any edited module YAML:
```bash
uv run linkml-validate -s src/ai_gene_review/schema/gene_review.yaml -C ModuleReview modules/<module>.yaml
uv run python -m ai_gene_review.validation.module_validator modules/<module>.yaml
just render-module modules/<module>.yaml
git diff --check
```
Check that each edited standalone module has a real decomposition. Prefer at
least two root `module.parts`; a root `variant_sets` decomposition is acceptable
only when it has multiple substantive alternatives. A single gene, enzyme, or
reaction should be recorded as pathway/gene curation instead of a module:
```bash
python3 - <<'PY' modules/<module>.yaml
import sys, yaml
for path in sys.argv[1:]:
data = yaml.safe_load(open(path)) or {}
module = data.get("module") or {}
parts = module.get("parts") or []
variants = sum(len(vs.get("variants") or []) for vs in module.get("variant_sets") or [])
if len(parts) < 2 and variants < 2:
raise SystemExit(f"{path}: standalone modules need at least two substantive parts or variants")
PY
```
If project markdown changed, render the touched project page:
```bash
uv run ai-gene-review render-projects projects/<path>.md -o pages/projects
```
If concrete gene reviews were edited or are central to the module claim, run:
```bash
just validate <species> <gene>
```
## Derived QC
Rendered module pages include derived QC from `ai_gene_review.module_qc`:
- recommended-field compliance;
- module deep-research provider detection;
- leaf nodes lacking representative members;
- gene-review completeness for UniProt groundings;
- advisory reaction chaining for `PRECEDES` connections.
Treat leaf-grounding warnings as modeling signals. Add a concrete UniProt
representative only when it truly helps orient an abstract family role; do not
add every organism mentioned in evidence just to improve a count.
## Cache Noise
Validation may update ontology cache files such as `cache/go/terms.csv`. Do not
include these incidental timestamp/cache changes in a module-curation PR unless
the task explicitly requires cache refresh. Remove them from the diff before
staging.
## PR Scope
A clean module-curation PR usually includes:
- the edited `modules/<module>.yaml`;
- rendered `pages/modules/<module>.html`;
- any touched project markdown and rendered page;
- generated deep-research files only if they were produced by an approved
provider command and are part of the requested work.
The PR body should state:
- what boundary/modeling decision changed;
- where terms were placed and why;
- which exemplars/PTNs/GO-CAMs/gene reviews ground the module;
- validation and render commands run;
- known warnings or follow-up issues.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |