This document defines Permea Core's initial attribution policy.
The goal is to recognize useful contributions, reviewer work, source dependencies, dataset/provenance work, derivative project maintenance, and paper-support work without confusing attribution with authorship, endorsement, validation, or release approval.
Attribution should be:
- accurate
- proportional
- public-safe
- respectful of confidentiality requests
- separate from authorship unless authorship criteria are met
- separate from endorsement unless endorsement is explicitly approved
- aligned with release and paper records
Attribution does not equal authorship. Attribution does not equal endorsement.
Contributes scripts, package code, tests, tooling, examples, or workflow improvements.
Improves docs, examples, templates, policy wording, checklists, or public-facing guidance.
Contributes benchmark contracts, task definitions, split policies, metric definitions, baseline designs, or benchmark review.
Contributes dataset source documentation, provenance manifests, release posture review, source-to-row mapping, license notes, or attribution requirements.
Dataset/provenance attribution does not mean dataset redistribution is permitted.
Improves run manifests, environment capture, rerun instructions, result artifact traceability, or reproducibility checks.
Reviews claims, scientific framing, reproducibility, benchmark contracts, dataset/provenance status, paper/source-to-claim alignment, or wet-lab protocol framing.
Review does not equal peer review, external validation, wet-lab validation, or clinical endorsement.
Coordinates review, triage, merge, policy enforcement, release readiness, or stewardship.
Maintainer attribution does not automatically imply release approval for every artifact.
Contributes protocol review, assay planning, experimental data, data interpretation, or partner review.
Wet-lab collaboration interest does not equal wet-lab validation.
Contributes through a university, nonprofit, lab, company, or partner organization.
Institutional attribution requires permission and should not imply institutional endorsement unless explicitly approved.
Maintains a Permea-aligned, official, or independent derivative project.
Permea-aligned does not mean Permea-validated.
Attribution records contribution or source credit.
Paper authorship requires paper-specific contribution, accountability, final approval, and manual decision under docs/contributors/AUTHORSHIP_POLICY.md.
Merged pull requests, issue comments, review notes, dataset suggestions, or acknowledgements do not automatically create paper authorship.
Attribution does not imply:
- endorsement
- validation
- official Permea approval
- public release approval
- dataset redistribution permission
- clinical or therapeutic support
- institutional endorsement
- partner endorsement
State endorsement separately only when explicitly approved and documented.
Derivative projects should attribute:
- Permea Core materials used
- specific release, tag, commit, schema, or benchmark contract used
- modified files or derived policies
- project status: official, Permea-aligned, or independent fork
- claim and release limitations
Derivative projects must not imply official Permea status, Permea validation, or release approval unless recorded.
Dataset and source attribution should record:
- dataset/source name
- source URL or citation
- source version or download date where available
- source license or terms where available
- required attribution wording where available
- redistribution status
- row-level release status
- derived artifact release status
- unresolved caveats
Do not claim dataset redistribution permission unless source/license approval is documented.
Paper and preprint attribution should align:
- manuscript author list
- acknowledgements
- contributor records
- reviewer credit
- dataset/source citations
- code and artifact citations
- release notes
- public repository docs
Paper-specific evidence packages should cite the relevant paper/preprint and the exact Permea Core release, tag, commit, or artifact used where appropriate.
Release notes may attribute:
- contributors
- reviewers
- maintainers
- benchmark/schema authors
- reproducibility contributors
- dataset/provenance reviewers
- derivative project maintainers
Release-note attribution should not imply endorsement, validation, authorship, or release approval beyond the recorded role.
Reviewer credit should follow docs/community/REVIEWER_CREDIT_POLICY.md.
Reviewer credit may appear as:
- GitHub review credit
- release-note credit
- public acknowledgement
- paper acknowledgement
- authorship consideration when authorship criteria are met
Reviewer input does not equal external validation.
Contributors or reviewers may request anonymous or confidential handling when:
- review is sensitive
- partner or institution constraints apply
- unpublished data or manuscripts are involved
- public naming could expose private information
Confidential attribution should be recorded privately where appropriate and summarized publicly only in public-safe terms.
Do not expose private correspondence, confidential partner information, or sensitive reviewer identities without permission.
Manual approval is required before:
- listing paper authorship
- converting attribution to authorship
- naming confidential contributors publicly
- using institutional affiliation in release-sensitive contexts
- implying endorsement
- naming contributors in partner-sensitive or wet-lab-sensitive materials
Attribution disputes should be escalated to maintainers and founder/manual approvers until a fuller dispute process exists.
This policy does not make legal conclusions about employment, IP ownership, trademark ownership, university policy, institutional policy, or authorship disputes.
Permea may use "AlphaFold for Delivery" as ambition or positioning only.
Permea must not claim AlphaFold-level performance, adoption, or standardization; completed wet-lab validation unless documented; clinical efficacy; universal delivery prediction; production-grade drug delivery platform status; or dataset redistribution permission without source/license approval.
Benchmark claims must remain scoped to dataset, split, metric, model, artifact, and evidence level.
Attribution does not equal authorship. Attribution does not equal endorsement. Archived does not mean validated. Permea-aligned does not mean Permea-validated.