@pikaa-ai/pikaa 0.3.1 → 0.3.4

package/skills/scientific-brainstorming/references/responsible_ai.md

@@ -1,220 +0,0 @@
-# Responsible AI in Research Ideation
-
-AI can supply prompts, reframings, counterarguments, or organizational help.
-It is not an expert panel, evidence source, author, ethics reviewer, or
-scientific decision maker. Capabilities and policies change; follow current
-institutional, funder, publisher, legal, and community requirements.
-
-## Default sequence
-
-1. **Classify information.** Decide whether the prompt would include personal,
-   patient, confidential, unpublished, proprietary, export-controlled,
-   security-sensitive, or otherwise restricted information.
-2. **Generate human ideas first.** Freeze an independent human-only round
-   before showing AI suggestions.
-3. **Define the AI role.** Examples: produce orthogonal questions, challenge an
-   assumption, list search terms, or reformat an already approved record.
-4. **Use only an approved tool and data class.** Do not assume a paid or
-   “private” interface satisfies institutional controls.
-5. **Capture provenance.** Record tool/model or service, date, purpose, material
-   prompt constraints, output IDs used, and human editor. Avoid storing
-   restricted prompt text in the session register.
-6. **Verify externally.** Check factual claims and every citation against
-   authoritative sources. Search for contradictory and null evidence.
-7. **Run a second independent human round.** Ask for ideas outside the AI's
-   frames and for harms or stakeholders the output omitted.
-8. **Disclose as required.** The accountable humans retain authorship,
-   responsibility, and final judgment.
-
-AI use is optional. The bundled CLIs make no network or LLM calls.
-
-## Suitable bounded roles
-
-- Generate alternative phrasings of a non-sensitive focal question.
-- Suggest dimensions for a morphological matrix, followed by human review.
-- Produce counterexamples or alternative mechanisms for registered ideas.
-- Identify ambiguous terms or missing assumptions.
-- Generate candidate search vocabulary, not references presented as real.
-- Convert an approved, non-sensitive record between formats.
-- Act as one disclosed adversarial prompt after human-first ideation.
-
-Avoid using AI to:
-
-- decide which scientific claim is true;
-- certify novelty, safety, ethics, legality, or regulatory compliance;
-- invent or complete missing data;
-- rank people, patients, communities, or protected groups;
-- replace stakeholder participation or domain expertise;
-- generate actionable harmful or dual-use procedures;
-- review confidential manuscripts, grants, or peer-review material in systems
-  where confidentiality is not assured.
-
-## Hallucination and source verification
-
-Generative systems can produce plausible but false claims, references,
-methods, statistics, and quotations. A 2023 study found fabricated and
-substantively erroneous bibliographic citations in outputs from the tested
-GPT-3.5 and GPT-4 versions; model-specific rates are not timeless estimates.
-
-For every AI-suggested source:
-
-1. locate the work in a trusted index or publisher site;
-2. match title, authors, venue, year, DOI, and version;
-3. read the relevant primary text;
-4. confirm the cited result, population, design, and limitations;
-5. record the stable source identifier;
-6. delete unsupported claims rather than laundering them as “AI suggested.”
-
-Never cite the model as evidence for a scientific claim.
-
-## Anchoring and homogenization
-
-AI output can anchor users on examples and compress a group's idea diversity.
-In a preregistered short-story experiment, access to GPT-4 ideas improved
-average evaluated creativity for some writers while making outputs more
-similar in aggregate. The task was short creative writing, not scientific
-ideation, so treat homogenization as a credible risk to test—not a universal
-effect size.
-
-Controls:
-
-- human-only generation before any AI output;
-- different participants receive no AI, or distinct prompt frames, when the
-  comparison is methodologically justified;
-- ask for mechanisms that contradict the AI's dominant frame;
-- compare assumptions, predictions, and causal structure—not only wording;
-- preserve pre-AI ideas and record which ideas changed after exposure;
-- include non-AI domain, methods, stakeholder, ethics, and safety perspectives;
-- do not infer independent support from many outputs of the same model.
-
-Multiple AI samples are correlated products of a system, not independent
-experts or replications.
-
-## Automation bias and false authority
-
-Fluent language, technical detail, and confident formatting are not evidence.
-To reduce deference:
-
-- hide model branding during idea review when feasible;
-- evaluate ideas against the same predeclared criteria;
-- require a human rationale and uncertainty statement;
-- assign a non-originating human challenger;
-- verify with primary evidence and domain experts;
-- retain a “no decision / insufficient evidence” outcome;
-- prohibit automatic advancement based only on an AI or matrix score.
-
-Do not ask an AI system to assign a probability it cannot calibrate and then
-treat the number as measured uncertainty.
-
-## Confidentiality, privacy, and intellectual property
-
-Do not submit the following to an external AI service unless an authorized
-policy and agreement explicitly permit that data class:
-
-- patient, participant, employee, student, or other personal information;
-- unpublished manuscripts, peer reviews, grants, invention disclosures, or
-  partner materials;
-- proprietary protocols, source code, compounds, sequences, or business data;
-- controlled unclassified, export-controlled, classified, or
-  security-sensitive information;
-- credentials, tokens, private links, or internal system details;
-- community-governed or Indigenous data outside agreed governance.
-
-Data minimization and abstraction are still required with an approved tool.
-Check retention, training use, access, location, deletion, audit, and incident
-terms. If the work cannot be safely abstracted, use an approved local/closed
-process or do not use AI.
-
-## Bias, representation, and participation
-
-AI output may reproduce gaps and stereotypes in training data and overrepresent
-well-indexed, English-language, high-resource perspectives. It cannot consent
-on behalf of affected communities.
-
-- Ask which populations, languages, geographies, disciplines, and negative
-  findings are missing.
-- Involve relevant people directly and compensate them where applicable.
-- Distinguish biological variables from social identities and avoid
-  essentialist mechanisms.
-- Examine whether a proposed measurement or intervention transfers across
-  settings.
-- Treat accessibility, equity, and distribution of benefits and burdens as
-  review criteria and possible gates.
-
-## Research integrity and disclosure
-
-Humans remain responsible for accuracy, attribution, originality, permissions,
-and the research record. AI systems should not be listed as authors. Record and
-disclose AI use at the level required by the institution, funder, venue, and
-applicable guidance.
-
-A useful internal disclosure includes:
-
-```text
-Tool/service and model or version (if exposed):
-Date used:
-Purpose:
-Information classification and approved environment:
-Human-first idea set frozen before use: yes/no
-Outputs retained or used:
-Verification performed:
-Material changes made by humans:
-Known limitations:
-```
-
-Disclosure does not cure inappropriate data sharing, plagiarism, fabricated
-citations, or unverified content.
-
-## Dual-use and misuse review
-
-AI can make technical ideation faster and more accessible. Screen both the
-research idea and the AI interaction for misuse potential.
-
-Escalate before generating operational detail when an idea could materially
-enable:
-
-- pathogen enhancement, immune evasion, host-range change, or harmful delivery;
-- synthesis, acquisition, concealment, scaling, or dissemination of hazardous
-  agents or toxins;
-- bypassing safety, monitoring, access, or security controls;
-- dangerous chemical, biological, cyber, autonomous, or surveillance
-  capability;
-- targeting vulnerable populations or critical systems.
-
-Use high-level risk framing while waiting for institutional biosafety,
-biosecurity, research-security, legal, ethics, or funding-agency guidance.
-Do not rely on a model's refusal behavior as a risk-management control.
-
-WHO's responsible life-sciences framework treats risk mitigation as a shared,
-multi-stakeholder responsibility. U.S. DURC/PEPP oversight has been under
-revision following the May 2025 executive order; verify current policy rather
-than copying a superseded threshold.
-
-## Incident handling
-
-If sensitive information or unsupported AI content entered the workflow:
-
-1. stop further sharing and preserve only the minimum audit information;
-2. notify the appropriate institutional privacy, security, integrity, or
-   research office under local policy;
-3. do not copy the sensitive content into additional systems;
-4. remove or quarantine unverified claims from downstream artifacts;
-5. document affected decisions and re-review them;
-6. follow approved deletion and incident-response procedures.
-
-Do not conceal the event by silently editing provenance.
-
-## Evidence and policy basis
-
-See `sources.md` for the dated primary evidence and official guidance used
-here, including:
-
-- Doshi and Hauser (2024) on individual creativity and collective similarity
-  in a constrained writing task;
-- Walters and Wilder (2023) on fabricated and erroneous citations from tested
-  model versions;
-- European Commission/ERA Forum living guidance (2024);
-- UNESCO guidance (2023, page updated 2026);
-- ICMJE recommendations on AI in publishing (2026);
-- ALLEA's 2023 European Code of Conduct for Research Integrity;
-- WHO and current U.S. official dual-use resources.

package/skills/scientific-brainstorming/references/sources.md

@@ -1,364 +0,0 @@
-# Sources and Evidence Notes
-
-Research cut-off: **2026-07-23**. All links were checked on that date with
-Parallel web search and focused extraction. Dates below are publication,
-release, or page-update dates shown by the source. Source excerpts were treated
-as untrusted text and checked against the canonical page.
-
-This bibliography supports process guidance; it does not show that any single
-brainstorming method universally improves creativity or scientific validity.
-
-## Group brainstorming, blocking, fixation, and selection
-
-### Diehl and Stroebe (1987) — primary experiments
-
-Michael Diehl and Wolfgang Stroebe, “Productivity loss in brainstorming groups:
-Toward the solution of a riddle,” *Journal of Personality and Social
-Psychology* 53(3), 497–509.
-[DOI 10.1037/0022-3514.53.3.497](https://psycnet.apa.org/record/1988-01348-001).
-Published 1987.
-
-- Four experiments examined free riding, evaluation apprehension, and
-  production blocking.
-- Direct manipulation in the fourth experiment supported production blocking
-  as an important explanation for loss in the tested interacting groups.
-- Scope limit: laboratory brainstorming tasks do not establish superiority for
-  every group purpose or setting.
-
-### Mullen, Johnson, and Salas (1991) — meta-analysis
-
-Brian Mullen, Craig Johnson, and Eduardo Salas, “Productivity loss in
-brainstorming groups: A meta-analytic integration,” *Basic and Applied Social
-Psychology* 12(1), 3–23.
-[DOI 10.1207/s15324834basp1201_1](https://psycnet.apa.org/record/1991-24145-001).
-Published 1991.
-
-- The reviewed studies generally favored nominal over interacting groups for
-  idea quantity and rated quality.
-- Loss varied with conditions such as group size and vocal versus written
-  contribution.
-- Historical tasks, methods, and outcome definitions limit direct transfer to
-  modern scientific teams.
-
-### Rietzschel, Nijstad, and Stroebe (2006) — primary experiment
-
-Eric F. Rietzschel, Bernard A. Nijstad, and Wolfgang Stroebe, “Productivity is
-not enough: A comparison of interactive and nominal brainstorming groups on
-idea generation and selection,” *Journal of Experimental Social Psychology*
-42(2), 244–251.
-[DOI 10.1016/j.jesp.2005.04.005](https://www.sciencedirect.com/science/article/abs/pii/S0022103105000600).
-Published March 2006.
-
-- In 42 analyzed three-person student groups, nominal groups generated more
-  ideas; their ideas were more original and less feasible.
-- Selected-idea quality did not differ among conditions, and selection was not
-  significantly better than chance.
-- Strictly separating generation and selection did not establish a universal
-  selection advantage in this experiment.
-
-### Rietzschel, Nijstad, and Stroebe (2010) — primary experiments
-
-Eric F. Rietzschel, Bernard A. Nijstad, and Wolfgang Stroebe, “The selection of
-creative ideas after individual idea generation: Choosing between creativity
-and impact,” *British Journal of Psychology* 101(1), 47–68.
-[DOI 10.1348/000712609X414204](https://pubmed.ncbi.nlm.nih.gov/19267959/).
-Published February 2010.
-
-- Explicit selection criteria improved selection on the named originality
-  dimension in the reported studies, with trade-offs in satisfaction or rated
-  effectiveness.
-- This supports explicit criteria, not a claim that a single composite score
-  identifies the scientifically best idea.
-
-### Smith, Ward, and Schumacher (1993) — primary experiments
-
-Steven M. Smith, Thomas B. Ward, and Jay S. Schumacher, “Constraining effects
-of examples in a creative generation task,” *Memory & Cognition* 21, 837–845.
-[DOI 10.3758/BF03202751](https://pubmed.ncbi.nlm.nih.gov/8289661/).
-Published November 1993.
-
-- Across three creative-generation experiments, participants exposed to
-  examples were more likely to reproduce example features.
-- The specific drawing/design tasks support an anchoring or fixation risk; they
-  do not quantify the effect for scientific ideation.
-
-### Esser (1998) — review
-
-James K. Esser, “Alive and Well after 25 Years: A Review of Groupthink
-Research,” *Organizational Behavior and Human Decision Processes* 73(2–3),
-116–141.
-[DOI 10.1006/obhd.1998.2758](https://pubmed.ncbi.nlm.nih.gov/9705799/).
-Published February 1998.
-
-- Reviews the evidence and evolution of Janis's groupthink model.
-- The broader literature includes substantial debate over the model and its
-  antecedents; this skill therefore uses “groupthink” as a risk prompt, not a
-  definitive diagnosis.
-
-## Nominal groups, Delphi, and structured elicitation
-
-### Van de Ven and Delbecq (1972) — early NGT paper
-
-Andrew H. Van de Ven and André L. Delbecq, “The nominal group as a research
-instrument for exploratory health studies,” *American Journal of Public
-Health* 62(3), 337–342.
-[DOI 10.2105/AJPH.62.3.337](https://pmc.ncbi.nlm.nih.gov/articles/PMC1530096/).
-Published March 1972.
-
-- Primary early account of the nominal-group approach in exploratory health
-  research.
-- Historical evidence should be combined with current, context-specific
-  reporting guidance.
-
-### Harb et al. (2021) — NGT scoping review
-
-Sami I. Harb et al., “Methodological options of the nominal group technique for
-survey item elicitation in health research: A scoping review,” *Journal of
-Clinical Epidemiology* 139, 140–148.
-[DOI 10.1016/j.jclinepi.2021.08.008](https://www.sciencedirect.com/science/article/abs/pii/S0895435621002535).
-Published November 2021.
-
-- Included 57 studies and identified 30 process decision points across five
-  broad stages.
-- Supports documenting actual choices and their rationale rather than treating
-  NGT as one invariant protocol.
-- Scope is health-survey item elicitation, not every use of NGT.
-
-### Jünger et al. (2017) — CREDES systematic review and guidance
-
-Saskia Jünger et al., “Guidance on Conducting and REporting DElphi Studies
-(CREDES) in palliative care,” *Palliative Medicine* 31(8), 684–706.
-[DOI 10.1177/0269216317690685](https://pubmed.ncbi.nlm.nih.gov/28190381/).
-E-published 13 February 2017.
-
-- A methodological systematic review of 30 palliative-care Delphi studies
-  found substantial variation in conduct, terminology, and reporting.
-- CREDES is reporting and conduct guidance developed in a palliative-care
-  context, not proof that Delphi consensus is correct.
-
-### RAND (2023) — official Delphi methods guidance
-
-Dmitry Khodyakov, Sean Grant, Jack Kroger, and Melissa Bauman,
-[*RAND Methodological Guidance for Conducting and Critically Appraising Delphi
-Panels*](https://www.rand.org/pubs/tools/TLA3082-1.html).
-RAND Corporation, 29 December 2023.
-[DOI 10.7249/TLA3082-1](https://doi.org/10.7249/TLA3082-1).
-
-- Defines Delphi as iterative, anonymous, structured group communication under
-  uncertainty and provides a design/appraisal tool.
-- RAND explicitly notes a lack of consistent methodological guidance across
-  the many Delphi variants.
-
-### EFSA (2014) — official expert-knowledge elicitation guidance
-
-European Food Safety Authority, “Guidance on Expert Knowledge Elicitation in
-Food and Feed Safety Risk Assessment,” *EFSA Journal* 12(6):3734.
-[DOI 10.2903/j.efsa.2014.3734](https://efsa.onlinelibrary.wiley.com/doi/10.2903/j.efsa.2014.3734).
-Adopted 22 May and published 19 June 2014.
-
-- Covers problem definition, preparation, expert/method selection, elicitation,
-  uncertainty, aggregation, and documentation.
-- Developed for EFSA food/feed risk assessment; apply domain-specific guidance
-  elsewhere.
-
-### Cooke, Mendel, and Thijs (1988) — structured judgment method
-
-Roger Cooke, Max Mendel, and Wim Thijs, “Calibration and information in expert
-resolution; a classical approach,” *Automatica* 24(1), 87–94.
-[DOI 10.1016/0005-1098(88)90011-8](https://www.sciencedirect.com/science/article/abs/pii/0005109888900118).
-Published 1988.
-
-- Presents calibration and information concepts for combining expert
-  judgments and an experiment with descriptive value.
-- Using a “calibration” label without the method's elicitation, seed questions,
-  and validation requirements would be misleading.
-
-## Rigor, reproducibility, and sex as a biological variable
-
-### NIH rigor and reproducibility guidance
-
-NIH Office of Extramural Research,
-[*Guidance: Rigor and Reproducibility in Grant Applications*](https://grants.nih.gov/policy-and-compliance/policy-topics/reproducibility/guidance).
-Page dated 16 October 2024.
-
-- Identifies rigor of prior research, rigorous experimental design, relevant
-  biological variables, and authentication of key resources.
-- This is grant guidance. It does not validate an idea or substitute for a
-  protocol, statistical review, or field-specific reporting standard.
-
-### NIH SABV policy and current ORWH guidance
-
-NIH Office of Research on Women's Health,
-[*Sex as a Biological Variable*](https://orwh.od.nih.gov/sex-as-biological-variable).
-Page last updated 17 October 2025.
-
-- NIH expects sex to be considered in design, analysis, and reporting for
-  vertebrate animal and human studies and requests strong justification for a
-  single-sex scope.
-- The page links the governing
-  [NOT-OD-15-102](https://grants.nih.gov/grants/guide/notice-files/not-od-15-102.html),
-  released 9 June 2015 and effective for relevant applications from 2016.
-- Scope is NIH policy; other funders and jurisdictions may differ.
-
-## Preregistration and open science
-
-### Nosek et al. (2018) — preregistration perspective
-
-Brian A. Nosek et al., “The preregistration revolution,” *Proceedings of the
-National Academy of Sciences* 115(11), 2600–2606.
-[DOI 10.1073/pnas.1708274114](https://pubmed.ncbi.nlm.nih.gov/29531091/).
-Published 13 March 2018.
-
-- Frames preregistration as a way to distinguish prediction from postdiction
-  by specifying questions and analyses before observing outcomes.
-- It is a methodological perspective, not evidence that registration alone
-  guarantees rigor or reproducibility.
-
-### Nosek et al. (2015) — TOP Guidelines
-
-Brian A. Nosek et al., “Promoting an open research culture,” *Science*
-348(6242), 1422–1425.
-[DOI 10.1126/science.aab2374](https://www.science.org/doi/10.1126/science.aab2374).
-Published 26 June 2015.
-
-- Proposes modular journal-policy standards for transparency and openness.
-- A policy framework does not itself establish that a specific study is valid
-  or ethically shareable.
-
-### COS/OSF — current operational guidance
-
-Center for Open Science,
-[*Preregistration*](https://www.cos.io/initiatives/prereg) and
-[*Preregistration: A Plan, Not a Prison*](https://www.cos.io/blog/preregistration-plan-not-prison).
-Current pages accessed 23 July 2026.
-
-- Explicitly values both exploratory and confirmatory research, recommends
-  transparent labeling, and explains how to report deviations.
-
-Center for Open Science,
-[*Registered Reports*](https://www.cos.io/initiatives/registered-reports).
-Current page accessed 23 July 2026.
-
-- Describes Stage 1 protocol review and in-principle acceptance before results,
-  with exploratory analyses reported separately.
-- COS states that Registered Reports are not a panacea for every field or
-  design.
-
-## Responsible AI and research integrity
-
-### Doshi and Hauser (2024) — primary randomized experiment
-
-Anil R. Doshi and Oliver P. Hauser, “Generative AI enhances individual
-creativity but reduces the collective diversity of novel content,” *Science
-Advances* 10(28):eadn5290.
-[DOI 10.1126/sciadv.adn5290](https://www.science.org/doi/10.1126/sciadv.adn5290).
-Published 12 July 2024.
-
-- Preregistered online experiment: 293 analyzed UK participants wrote
-  eight-sentence stories with no, one, or up to five GPT-4 ideas.
-- AI access improved average evaluated novelty/usefulness, especially for
-  lower-scoring writers, while outputs became more similar in aggregate.
-- Constrained short-story writing is not scientific ideation; the study
-  supports a risk hypothesis, not a universal creativity claim.
-
-### Walters and Wilder (2023) — primary citation audit
-
-William H. Walters and Esther Isabelle Wilder, “Fabrication and errors in the
-bibliographic citations generated by ChatGPT,” *Scientific Reports* 13.
-[DOI 10.1038/s41598-023-41032-5](https://www.nature.com/articles/s41598-023-41032-5).
-Published 7 September 2023.
-
-- Audited 84 texts produced in April 2023 by tested GPT-3.5 and GPT-4 versions
-  and documented fabricated and substantive citation errors.
-- Results are model-, prompt-, task-, and date-specific; they justify direct
-  source verification rather than a timeless error rate.
-
-### European Commission / ERA Forum (2024) — official living guidance
-
-European Commission Directorate-General for Research and Innovation,
-[*Guidelines on the responsible use of generative AI in
-research*](https://research-and-innovation.ec.europa.eu/news/all-research-and-innovation-news/guidelines-responsible-use-generative-ai-research-developed-european-research-area-forum-2024-03-20_en).
-Published 20 March 2024.
-
-- Emphasizes transparency, responsibility, privacy, confidentiality,
-  intellectual property, bias awareness, and avoiding AI in sensitive
-  evaluation activities when confidentiality is not assured.
-- The Commission describes the guidance as living and subject to updates.
-
-### UNESCO (2023; updated 2026) — official guidance
-
-Fengchun Miao and Wayne Holmes,
-[*Guidance for generative AI in education and research*](https://www.unesco.org/en/articles/guidance-generative-ai-education-and-research).
-Published 7 September 2023; page last updated 16 January 2026.
-
-- Uses a human-centered framing and addresses privacy, ethical validation,
-  inclusion, equity, and institutional capacity.
-- Much of the document concerns education; apply research-specific provisions
-  with local policy.
-
-### ICMJE (2026) — official publishing recommendations
-
-International Committee of Medical Journal Editors,
-[*Use of Artificial Intelligence in Publishing*](https://www.icmje.org/recommendations/browse/artificial-intelligence).
-Recommendations page dated 2026.
-
-- Humans remain responsible for accuracy, attribution, permissions,
-  confidentiality, and disclosure; AI tools are not authors.
-- Publishing guidance does not replace institutional research policy.
-
-### ALLEA (2023) — research-integrity code
-
-ALLEA,
-[*The European Code of Conduct for Research Integrity, 2023 Revised
-Edition*](https://allea.org/portfolio-item/european-code-of-conduct-2023).
-[DOI 10.26356/ECOC](https://doi.org/10.26356/ECOC). Published 2023.
-
-- Cross-disciplinary self-regulatory framework recognized as a reference for
-  EU-funded research.
-- General principles require translation into institutional and
-  discipline-specific procedures.
-
-## Dual-use and responsible life sciences
-
-### WHO global framework and current program
-
-World Health Organization,
-[*Ensuring responsible use of life sciences research*](https://www.who.int/activities/ensuring-responsible-use-of-life-sciences-research).
-Program page accessed 23 July 2026; it indexes the 13 September 2022 Global
-Guidance Framework and materials through a 15 June 2026 meeting report.
-
-- Frames dual-use risk mitigation as individual and collective,
-  multi-stakeholder work across life sciences and relevant emerging
-  technologies.
-- Global guidance must be applied with jurisdictional and institutional rules.
-
-### Current U.S. policy-status notice
-
-U.S. Department of Health and Human Services, Administration for Strategic
-Preparedness and Response,
-[*Dual Use Research of Concern Oversight Policy Framework*](https://aspr.hhs.gov/S3/Pages/Dual-Use-Research-of-Concern-Oversight-Policy-Framework.aspx).
-Current-status notice accessed 23 July 2026.
-
-- The page states that federal departments and agencies will revise or replace
-  the 2024 DURC/PEPP policy in response to the 5 May 2025 executive order and
-  that the page will be updated when revised policy is available.
-- Therefore this skill does not present the 2024 policy as a stable current
-  checklist; researchers should consult their institution and current agency
-  materials.
-
-## Research method
-
-Parallel CLI searches used focused academic and official-domain queries for:
-
-- interactive versus nominal brainstorming and productivity blocking;
-- NGT, Delphi, and structured expert elicitation;
-- divergent/convergent idea generation and selection;
-- NIH rigor, reproducibility, and SABV;
-- preregistration, Registered Reports, and open-science guidance;
-- responsible AI, hallucinated citations, homogenization, confidentiality,
-  research integrity, and disclosure;
-- WHO and U.S. dual-use policy status.
-
-Canonical sources were then fetched with `parallel-cli extract` using focused
-objectives for bibliographic metadata, methods, findings, limitations, and
-current policy status. No research JSON artifacts were saved in the skill.