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/research-and-grounding.md
View on GitHub# Research And Grounding
## Module Research
Use module-level deep research to establish pathway boundaries, missing steps,
candidate members, and known variants. When the task is species/pathway
satisfiability, research the combination of module + pathway/accession + taxon
or species, not only the generic pathway.
Generated provider files must be produced by the configured commands. Do not
write original prose and name it `*-deep-research-<provider>.md`. If provider
research is unavailable or manual triage is needed, write notes in a normal
project/gene/module notes file or use a clearly manual filename.
Known commands from `modules/README.md`:
```bash
just module-deep-research-perplexity peroxisome-lifecycle
just module-deep-research-openai modules/gluconeogenesis.yaml
```
If a different provider is wired in locally, follow the repo's just recipes and
record the provider-specific output path in evidence.
## Gene Review Integration
For concrete UniProt members:
- ensure the corresponding gene review exists when the module relies on that
member;
- read the gene review's accepted core functions, non-core context, proposed
terms, and notes;
- do not use module curation to silently override a gene review; record
disagreements as knowledge gaps or follow-up questions;
- run `just fetch-gene <species> <gene>` only when creating a missing review
scaffold is part of the task.
For Pseudomonas putida KT2440, the species directory is `PSEPK`.
## Identifier Discipline
- Never guess GO, UniProt, PANTHER, PTN, Rhea, ChEBI, PMID, taxon, or gene
identifiers.
- Prefer local derived files first: gene `*-uniprot.txt`, `*-goa.tsv`,
`interpro/panther/**`, `gocams/index.tsv`, cached GO-CAMs, and cached
publications.
- If local files are insufficient and the identifier may have changed, look it
up from an authoritative source.
- For GO term placement, check the aspect: MF terms belong in annoton
`function`, BP terms in module/process concepts or annoton `processes`, and CC
terms in context or locations.
## Evidence Placement
- Put top-level `evidence` on the `ModuleReview` for sources supporting the
overall module boundary.
- Put node/part/annoton evidence as close as possible to the claim it supports.
- Use `notes` for curator rationale and caveats; use `knowledge_gaps` for open
biological or curation questions.
- Keep project-history commentary in `projects/**`, not in reusable module
descriptions.
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 |