All skills
imbad0202 avatar

/deep-research

@c5c1b45

Universal deep research agent team. 13-agent pipeline for rigorous academic research on any topic. 8 modes: full research, quick brief, paper review, lit-review, fact-check, three-way literature scan, Socratic guided research dialogue, and systematic review with optional meta-analysis. Covers research question formulation, Socratic mentoring, methodology design, systematic literature search, source verification, cross-source synthesis, risk of bias assessment, meta-analysis, APA 7.0 report compilation, editorial review, devil's advocate challenges, ethics review, and post-research literature monitoring. Triggers on: research, deep research, literature review, systematic review, meta-analysis, PRISMA, evidence synthesis, fact-check, WHY HOW WHAT papers, 3W literature scan, guide my research, help me think through, 研究, 深度研究, 文獻回顧, 文獻探討, 系統性回顧, 後設分析, 事實查核, 三段式文獻掃描, 引導我的研究, 幫我釐清, 幫我想想, 我不確定要研究什麼, 研究方向, 研究主題, 심층 연구, 문헌 조사, 체계적 문헌고찰, 메타분석, 사실 확인, 연구 방향을 잡아줘, 연구 주제 정하는 것을 도와줘, revisión de literatura, metaanálisis

Use this Skill: https://skilld.dev/gh/imbad0202/academic-research-skills/deep-research

This session only. Nothing lands on disk.

referencesirb_decision_tree.md

≈4.9k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Human-Subjects Authority Context Navigation Aid — Institutional Determination Required

Purpose

This reference helps the ethics_review_agent and research_architect_agent collect protocol facts, bind already-selected authority profiles, and prepare questions for the responsible institution. It is a navigation aid, not an authority, authorization tool, or review-pathway classifier.

The controlling local contract is shared/references/human_subjects_authority_protocol.md; the only curated authority rows available to this aid are in shared/human_subjects_authority_registry.json. Every profile in that registry is a bounded_subset, not a complete statement of any jurisdiction's law or an institution's policy.

When a caller explicitly requests candidate rule-trace display, the derived handoff is governed by shared/references/review_pathway_rule_trace_protocol.md. This file never supplies candidate names or unanchored predicates to that handoff.

Authority boundary (#665/#666/#680): Mixed-jurisdiction pathway mappings are not part of this reference. ARS output must use institutional determination required; applicability, requirements, timing, and authorization remain with the responsible institution and an exactly selected authority profile. Missing or unknown facts remain unresolved and never become a review determination.


1. Portable Fact Collection — No Pathway Inference

Collect facts without mapping them to Exempt, Expedited, Full Board, or a similarly named local pathway. Those labels are authority- and institution-specific; vulnerability, sensitivity, deception, identifiability, and estimated risk are inputs for human assessment, not automatic pathway rules.

Start
│
├── Record whether the activity involves human subjects
│   └── If unknown, preserve unknown and ask the responsible institution
│
├── Record each candidate review-ethics and data-protection authority
│   └── Do not infer jurisdiction from geography, funder, language, or topic
│
├── Record data, interaction, recruitment, population, and risk facts
│   └── Separate observed facts from researcher assumptions
│
├── Select exact #666 profile id/version/digest pointers only when confirmed
│   └── Leave an axis unresolved when no matching bounded profile is available
│
└── Send facts and unresolved questions to the responsible institution
    └── Output pathway: institutional determination required

1.1 Exact #666 declared facts

For a #666 context, record true, false, or unknown against these exact catalogue identifiers. A missing value is unknown, not false.

Exact fact id Question to resolve
activity.human_subjects Has the author confirmed that the activity involves human subjects?
support.us_hhs_or_common_rule Has the author identified a U.S. HHS/Common Rule scope basis?
scope.tw_hsra Has the author identified Taiwan's Human Subjects Research Act as a review authority?
scope.eu_gdpr Has the author identified the GDPR as a data-protection authority?
processing.personal_data Within the selected GDPR scope, does the activity process personal data?
data.special_category Does that processing include special-category data?
data.collected_from_subject Is the personal data obtained from the data subject rather than elsewhere?
purpose.scientific_research Does the processing purpose include scientific research?
gdpr.member_state_research_law_identified Has a specific applicable Union or Member State research-derogation law been identified?

1.2 Additional portable planning facts

The following are illustrative, unprofiled planning facts. They preserve useful protocol detail but are not predicates or requirement rows in the current #666 registry:

  • source and provenance of each dataset or specimen;
  • direct identifiers, quasi-identifiers, and linkage/re-link keys;
  • IP addresses, cookies, device identifiers, timestamps, and other platform metadata actually collected or accessible;
  • direct interaction, intervention, observation, recordings, and follow-up;
  • recruitment channel, compensation, dependency or power relationships, age, decision-making capacity, language, and accessibility needs;
  • potentially sensitive topics, deception or incomplete disclosure, and the physical, psychological, social, economic, informational, or group risks the team has identified;
  • risk mitigations, access controls, recipients, transfers, retention, destruction, reuse, withdrawal handling, and incident response;
  • institution-, funder-, community-, sector-, or profession-specific rules and the office or person responsible for interpreting them.

Do not convert any single fact or combination in this list into a pathway, clearance, readiness, compliance, or authorization result.

1.3 Derived #669 candidate rule trace

The #669 runtime may display only an explicit, replay-bound request that partitions every selected-profile pathway_trace requirement. Candidate labels come from named institutional material or an author-declared question, never from this reference. Both authority axes are accounted for, while alternatives remain within one exact profile; a GDPR/data-protection route is never presented as an IRB route.

Every emitted row backpoints to an exact requirement and authority anchor in section 2. An unknown fact stays in Unresolved predicates with the exact authoritative_decision_maker.role_id. No selected profile means no candidate rows and an exact JURISDICTION_UNRESOLVED halt. The only result is institutional determination required.

Consumers must receive confirmation that both validate_review_pathway_rule_trace(...) and the surface-scoped check_review_pathway_output.py lint succeeded against the exact named files. The trace is display-only and cannot update readiness, authorization, an action, a verdict, a checkpoint, or any workflow state.


2. Exact Bounded Requirement Pointers

Use a row below for action planning only after the matching profile has been selected by exact id/version/digest and the #666 resolved artifact permits profile-dependent use. The separately replayed #669 display-only trace may preserve a requirement-level unknown under its narrow protocol exception, but that row cannot become an action or result. The link is the row's primary-source anchor; the requirement id is the navigation pointer. The table is deliberately exhaustive only for the current bounded registry, not for the underlying authorities.

2.1 Review-ethics axis

Exact requirement id Exact provision Obligated actor Consumer scopes Bounded subject Primary-source anchor
us.45cfr46.107.irb-composition 45 CFR 46.107 institution_or_irb ["committee_governance", "pathway_trace"] IRB membership and composition 45 CFR 46.107
us.45cfr46.116.informed-consent 45 CFR 46.116 investigator_and_irb ["participant_information", "submission_packet", "pathway_trace"] General informed-consent requirements 45 CFR 46.116
tw.hsra.article-7.committee-composition 人體研究法第 7 條 research_institution_or_review_committee ["committee_governance", "pathway_trace"] Review-committee composition 人體研究法第 7 條(全國法規資料庫中文控制文本)
tw.hsra.article-14.consent-information 人體研究法第 14 條 principal_investigator ["participant_information", "submission_packet", "pathway_trace"] Information explained before consent 人體研究法第 14 條(全國法規資料庫中文控制文本)

Committee-composition rows belong to their stated institutional/committee actors. This aid must not turn them into investigator submission-packet requirements. The U.S. and Taiwan rows remain parallel and retain their own vocabulary; neither is a translation of the other.

2.2 Data-protection axis

Exact requirement id Exact provision Obligated actor Consumer scopes Bounded subject Primary-source anchor
eu.gdpr.article-6.lawful-basis GDPR Article 6 controller ["data_governance", "pathway_trace"] Identify and document an applicable Article 6 basis; the profile does not select one GDPR Article 6
eu.gdpr.article-9.special-category-research GDPR Article 9 controller ["data_governance", "pathway_trace"] Identify an applicable Article 9 condition and safeguards; the profile does not select the route GDPR Article 9
eu.gdpr.article-13.direct-collection-information GDPR Article 13 controller ["participant_information", "data_governance", "pathway_trace"] Information when data is obtained from the data subject GDPR Article 13
eu.gdpr.article-14.indirect-collection-information GDPR Article 14 controller ["participant_information", "data_governance", "pathway_trace"] Information when data is obtained elsewhere, subject to a separately established exception GDPR Article 14
eu.gdpr.article-89.1.research-safeguards GDPR Article 89(1) controller ["data_governance", "pathway_trace"] Scientific-research safeguards, including assessment of data minimisation and pseudonymisation GDPR Article 89(1)
eu.gdpr.article-89.2.member-state-derogation GDPR Article 89(2) controller ["data_governance", "pathway_trace"] A claimed research derogation must point to an identified applicable Union or Member State law GDPR Article 89(2)

The GDPR rows do not determine whether the GDPR applies, select a lawful basis or Article 9 condition, identify a Member State law, or authorize processing.


3. Illustrative Country and Institution Planning

3.1 Taiwan institutional-workflow example — illustrative and unprofiled

For a project connected with Taiwan, a useful planning conversation may ask:

  1. Which institution or review committee owns the determination?
  2. Which local submission system and current forms does it use?
  3. Which protocol, participant-information, recruitment, instrument, training, data-management, or other materials does that institution request?
  4. Which activities, if any, must wait for a documented institutional decision?
  5. What current turnaround estimate, amendment process, continuing-reporting schedule, and closure process does the institution state?

These are questions, not Taiwan legal requirements. The current #666 Taiwan profile covers only tw.hsra.article-7.committee-composition and tw.hsra.article-14.consent-information. It does not profile NSTC or Ministry of Education funding rules, the Personal Data Protection Act, clinical-trial or biobank rules, training credentials, submission platforms, timelines, periodic reports, or a local pathway taxonomy. Record any such item as unprofiled and obtain the responsible authority's current instruction. Do not rename a Taiwan process with U.S. pathway vocabulary.

3.2 Higher-education scenarios — illustrative fact prompts

Scenario Portable facts to collect; no determination implied
Public statistics analysis source, publication/access terms, record-level identifiability, linkage with other data, and institutional policy
Institutional research data provenance, authorization, direct/quasi-identifiers, re-link key, access roles, purpose, and data source
Student survey or interview recruitment and power relationship, voluntariness, direct identifiers, platform metadata, topic, recording, and withdrawal handling
Teaching intervention allocation, effect on grades/rights/access, alternatives, instructor-researcher roles, risk, and data collection
Mental-health or other sensitive-topic study data categories, foreseeable harms, disclosure limits, response plan, and—only if the GDPR profile is selected—whether data.special_category is true
Learning portfolio or classroom observation expectation of privacy, recorded identifiers, notice/interaction, recording, secondary use, and access
Career tracking or longitudinal follow-up contact data, re-link key, follow-up period, recipients, retention, and withdrawal feasibility
Cross-border or multi-institution study each review-ethics and data-protection authority, lead/participating institution roles, transfers, and unresolved conflicts

Public availability, anonymity, minimal risk, vulnerability, or sensitivity does not by itself assign a pathway in this reference.


4. Participant-Information and Consent Preparation

4.1 Authority-bound coverage

There is no universal consent-element checklist in this aid. When the applicable row resolves true, use its exact pointer and full registry row:

  • us.45cfr46.116.informed-consent for the bounded U.S. Common Rule consent row;
  • tw.hsra.article-14.consent-information for the bounded Taiwan Article 14 information row;
  • eu.gdpr.article-13.direct-collection-information when the selected GDPR context says personal data is obtained from the data subject;
  • eu.gdpr.article-14.indirect-collection-information when the selected GDPR context says personal data is obtained elsewhere.

Do not merge these rows into one universal form. If an authority, element, waiver/exception, special population, sector, or institutional rule is not in the registry, label it unprofiled and seek the responsible authority's current requirements.

4.2 Portable drafting inventory — illustrative, non-authoritative

The following inventory is useful for drafting and gap questions, but it is not a list of legally required elements:

  • project title; responsible institution, investigator, funder, and contacts;
  • purpose, procedures, duration, recordings, alternatives, and participant tasks;
  • foreseeable risks/discomforts, possible benefits, compensation, and remedies;
  • voluntary-choice and withdrawal statements appropriate to the actual design;
  • data categories and sources; collection method; recipients and access roles;
  • platform metadata, identifiers, linkage, security, retention, destruction, future use, sharing, transfer, publication, and commercial interests;
  • complaint, rights, or authority contacts supplied by the responsible regime;
  • signature, assent, representative, or electronic-interaction fields requested by the responsible institution.

4.3 Online-survey disclosure accuracy

Use this portable drafting convention: disclose what the selected survey platform, institution, and research team actually collect or can access. Verify settings and data exports before saying that an IP address, cookie, device identifier, timestamp, contact field, or other metadata is not recorded. State the collector, purpose, access roles/recipients, retention, and deletion or masking behavior where known; preserve unknown where it has not been verified.

An I agree control records a participant interaction. This aid does not decide whether that interaction satisfies consent, signature, documentation, or waiver requirements under a selected authority or institutional process.

4.4 Drafting worksheet — illustrative and unprofiled until bound

Participant Information / Consent Drafting Worksheet

Selected profile pointer(s): [profile id / version / digest]
Applicable requirement pointer(s): [exact requirement ids]
Unprofiled authorities or institutional rules: [list / unknown]

Project and responsible actors: [                              ]
Purpose, procedures, duration, and recordings: [               ]
Risks, benefits, compensation, and remedies: [                 ]
Choice, withdrawal, and limits on withdrawal/deletion: [       ]
Data and metadata actually collected (including platform data): [ ]
Access, recipients, transfers, security, and incident handling: [ ]
Retention, destruction, future use, sharing, and publication: [  ]
Contacts and institution-requested acknowledgement fields: [     ]

Unresolved questions for the responsible institution: [          ]
Review pathway: institutional determination required

5. Data-Handling Technical Planning

Use shared/references/irb_terminology_glossary.md for explicitly convention-bound drafting terms. A technical transformation does not, by itself, establish an authority-defined legal status.

5.1 Illustrative techniques

Technique Technical planning purpose; legal effect not asserted
Identifier removal or transformation reduce direct-identification exposure
Pseudonym/code assignment separate working records from additional identifying information while preserving a controlled link
Generalisation or aggregation reduce precision and small-cell disclosure risk
Masking limit visible portions of a field for a stated use
Re-identification risk assessment test plausible combinations and auxiliary-data attacks under a named method or threshold

5.2 Illustrative risk and control prompts

  • Can small groups, quotations, rare attributes, dates, locations, or institutional characteristics identify a person or community?
  • Can other datasets be linked to the planned release or analysis environment?
  • Who holds each identifier or re-link key, and who can access both sides?
  • Do storage, encryption, logging, export, backup, retention, and destruction controls match the participant-facing disclosure?
  • Has the team tested the actual platform export for IP addresses and other metadata rather than relying on a default-setting assumption?

Researcher Self-Check

Before handing the protocol to the responsible institution:

  1. Have both authority axes been addressed with exact selected profiles or a visible unresolved state?
  2. Are confirmed, false, and unknown facts kept distinct without jurisdiction inference?
  3. Does every authority-bound statement point to an exact #666 requirement id and its selected profile?
  4. Are actor and consumer boundaries preserved, especially for committee-governance rows?
  5. Are country, sector, community, funder, and institutional examples outside the registry labelled illustrative or unprofiled?
  6. Does survey/privacy wording accurately describe actual identifier and metadata handling?
  7. Are vulnerability, sensitivity, deception, identifiability, and risk treated as facts rather than automatic pathways?
  8. Is the pathway still institutional determination required, with timing and authorization obtained separately from the responsible institution?

Human-subjects boundary: This navigation aid does not authorize recruitment, consent, access to identifiable data, intervention, or data collection.

Source: SKILL.md on GitHub

No alerts1d5 checks · Risk SAFE
  • Gen Agent Trust Hub1d

    The deep-research skill is a highly sophisticated research orchestration system that includes robust internal security instructions to prevent prompt injection from untrusted data. It leverages well-known scholarly APIs for data verification and uses local scripts for deterministic validation of research artifacts.

  • Socket1d

    No alerts

  • Snyk1d

    Risk: LOW · No issues

  • Runlayer6mo

    7/41 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at c5c1b45. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 29 minutes ago.

Activeupdated 2 days ago
Other metadata
metadata
{
  "version": "2.12.1",
  "last_updated": "2026-08-15",
  "status": "active",
  "data_access_level": "raw",
  "task_type": "open-ended",
  "related_skills": [
    "academic-paper",
    "academic-pipeline"
  ]
}

README badge

README badge for imbad0202/academic-research-skills/deep-research