How this knowledge base is built
A transparent method matters more than a confident tone. This page names who is accountable, what AI assists with, what counts as evidence and when guidance must be reviewed.
- Publisher and accountable editor
- Molt & Moss, a one-person California Neocaridina breeder and publisher
- Independent peer review
- None claimed
- Structured evidence records
- 148
- Supporting authorities
- 29
- Named open questions
- 16
- Evidence registry reviewed
- August 12, 2026
- Next scheduled review
- February 12, 2027
- Editorial policy
- Version 1.0, effective August 12, 2026
Scope and accountability
The biological scope is Neocaridina. Aquarium-system methods, professional guidance and adjacent taxa appear only when their relationship is labelled. Molt & Moss accepts responsibility for the selection, interpretation and wording it publishes. No university, laboratory, veterinarian, journal or independent reviewer is implied.
This is educational husbandry material, not veterinary diagnosis, treatment authorization or a promise that one home aquarium will reproduce a study result.
How AI is used
AI tools assist this project. They may help discover candidate sources, organize notes, extract structured fields, draft explanations, write software and generate consistency tests. They are not authors, peer reviewers, experts or sources.
- A model answer, summary or citation is never evidence merely because it sounds plausible.
- A public claim must trace to the source registry, a separately labelled authority or an explicitly bounded house record.
- When full methods are unavailable, the item remains a candidate or full-text need and cannot support a detailed claim.
- Automated tests can detect missing fields, broken boundaries and inconsistent routes. They cannot prove that a scientific interpretation is true.
- The application does not send private guide feedback, questions, observations or evidence challenges to an AI model. Any future editorial use requires a separate privacy decision.
- A synthetic image must be labelled and cannot serve as evidence of a specimen, grade, lesion, organism or product received.
The evidence path
- Discover. Preserve the exact query, date and candidate decision rather than treating a search result as evidence.
- Retrieve. Prefer the complete primary article, accepted manuscript or authoritative record. Do not reconstruct hidden methods from an abstract.
- Structure. Record taxon, animals, experimental unit, comparison, duration, endpoint, result and material limitations.
- Bound. State what the source supports and the nearest tempting claim it does not establish.
- Synthesize. Write the practical guide from the bounded records, keeping controlled findings, inference and house practice separate.
- Review. Run claim-scope and consistency checks, expose the source trail and record meaningful additions, corrections and narrowed wording.
- Correct. Accept exact evidence challenges, publish finite outcomes and carry accepted changes into the guide and public history.
Source roles are not interchangeable
| Source role | Useful for | Does not automatically establish |
|---|---|---|
| Direct controlled Neocaridina experiment | The tested treatment, unit, duration and endpoint | A universal home optimum, diagnosis or safe dose |
| Field observation or taxonomy | Occurrence, anatomy, identity and association under stated methods | Home-aquarium causation or care targets |
| Review or synthesis | Scope, prior literature and named uncertainty | A new independent animal experiment |
| Adjacent taxon or aquarium system | A labelled mechanism, method or system property | A direct Neocaridina biological response |
| Professional, regulatory or manufacturer authority | Its exact method, rule, product instructions or professional boundary | Independent species evidence beyond that role |
| House record or community observation | A reproducible local observation and a question worth testing | A controlled study, pooled prevalence or universal rule |
What the review date means
The shared evidence-review date means the structured registry and its public claim boundaries were reconciled as a corpus. It is not the publication date of every source, proof of independent replication or a claim that every sentence became newly true that day.
A scheduled review occurs at least every six months. A credible correction, consequential new study, taxonomic change, changed authority or detected source conflict triggers earlier review. Cosmetic edits do not manufacture a fresh scientific date.
Commercial conflict
Molt & Moss intends to sell shrimp and supplies while publishing care guidance. A seller is not an independent certifier of its own products. Product benefit, sales, affiliate relationships, free material and contributor relationships cannot substitute for evidence and must be disclosed when material.
Audit or correct the work
- Separate locally verified capabilities from production, community and independent proof still required.
- Use the independent-review packet to audit consequential claims, evidence paths and conflicts.
- Apply source-role-specific critical-appraisal questions without a quality score.
- Inspect the structured evidence, supporting authorities, gaps and change history.
- Compare direct evidence, controlled studies and open questions by keeper topic.
- Audit the complete linked evidence dossier for any detailed guide.
- Inspect or reuse the versioned machine-readable knowledge export with its JSON Schema.
- Challenge an exact claim with a stable source.
- Ask a keeper question the current library cannot route.
- Compare every community contribution and evidence ceiling.