| name | cognitive-frameworks |
| description | Use before major research proposals, manuscripts, reports, methodology sections, literature reviews, grants, or stakeholder-facing academic/professional writing to make claims, gaps, warrants, boundaries, and section rhetoric explicit before drafting. |
Cognitive Frameworks
Use this skill before substantial research writing or revision when the output needs an argument, not just polished wording.
Trigger
Use for:
- proposal, manuscript, report, grant, protocol, thesis, dissertation, or chapter planning;
- literature review, methodology, introduction, rationale, discussion, recommendation, or implications sections;
- supervisor, PI, client, reviewer, examiner, funder, or stakeholder-facing drafts;
- research-gap, problem-statement, contribution, or design-rationale decisions;
- major revisions where the current text feels shallow, unfocused, or under-justified.
Do not use for short factual answers, file listing, simple formatting, or routine maintenance notes unless the user asks for deeper reasoning.
Required Inputs
Before using this skill, check:
AGENTS.md
PROJECT_AGENT_PREFERENCES.md
RESEARCH_PROJECT_BRIEF.md if present, otherwise RESEARCH_PROJECT_BRIEF_TEMPLATE.md
- relevant source files or source notes
knowledge-base/SOURCE_READINESS_MATRIX.md for citation-heavy work
If project facts or sources are missing, mark them TO CONFIRM.
Core Output
Create a short cognitive protocol before drafting:
Cognitive protocol:
- Section type:
- Main claim:
- Gap/problem type:
- Evidence base:
- Warrant:
- Boundary:
- Rhetorical plan:
- Risks:
- Next drafting move:
The protocol can be a temporary note, a checkpoint file, or a section inside a planning document.
Use research-wiki/CLAIM_LEDGER_LITE_PROTOCOL.md when formal claims, stakeholder-facing claims, source-readiness upgrades, or citation claim-support audits need stable claim_id boundaries. cognitive-frameworks prepares candidate claims, warrants, limitations, and concept contracts; dissertation-citation-audit owns the ledger check.
Gap / Problem Type Taxonomy
Use one or more of these allowed categories. Do not write a vague gap statement without naming the type.
| Type | Use When | Signal |
|---|
| Conceptual gap | Existing concepts do not clearly explain the target issue | terms are broad, contested, or under-specified |
| Empirical/context gap | A population, site, setting, or case is under-examined | evidence exists elsewhere but not in this context |
| Methodological gap | Existing studies rely on limited methods or miss a useful form of evidence | surveys only, no interviews, weak design input, limited triangulation |
| Implementation/adoption boundary | Existing work discusses value but not practical conditions for use | barriers, acceptability, workload, trust, adoption, governance |
| Ethics/governance gap | Existing work under-specifies risk, accountability, privacy, autonomy, or consent | compliance or responsible-use assumptions are unclear |
| Policy/practice gap | Evidence does not translate cleanly into guidance for practitioners or organisations | unclear actions, roles, or implementation pathway |
| Evidence synthesis gap | The issue is fragmented across literature areas | relevant work exists but is not connected for this project |
| Technical/design gap | A tool, system, workflow, or interface requirement is under-specified | design boundaries, feature requirements, failure modes |
If none fits, write Gap/problem type: TO CONFIRM and explain what source review is needed.
Claim-Evidence-Warrant Map
Every important claim should have:
- Claim: the sentence or judgement the document needs to defend.
- Evidence: source material, data, policy, user requirement, or project fact that supports the claim.
- Warrant: the reasoning that explains why the evidence supports the claim.
- Boundary: what the evidence does not prove.
Weak warrant examples:
- "This is supported by the literature."
- "This is important."
- "This shows the gap."
Stronger warrant pattern:
Because [evidence] shows [specific condition], it is reasonable to claim [claim] within [boundary].
Warrant Quality Test
Before drafting, test the warrant:
- Does the evidence actually lead to the claim, or only relate to the topic?
- What assumption connects the evidence to the claim?
- What counterexample would weaken this claim?
- Is the claim narrower than the evidence can support?
- Does the sentence need a source, a caveat, or
TO CONFIRM?
If the warrant fails, narrow the claim or return to source review.
Section Rhetorical Moves
Choose a section type and follow the relevant moves in:
references/research_rhetorical_moves.md
references/argumentation_reasoning_framework.md
Minimum requirement:
- Introduction/rationale: establish issue, locate literature, define gap/problem type, state project response.
- Literature review: group sources by argument job, compare positions, identify usable evidence and boundaries.
- Methodology/methods: connect research question, data, method, ethics/compliance, and limitations.
- Findings/results: separate observed evidence from interpretation.
- Discussion/implications: show what the findings mean, where they apply, and what remains limited.
- Proposal/grant/report recommendation: show need, feasible action, evidence basis, risk, and decision point.
Boundary
This skill improves reasoning discipline. It does not:
- verify citations;
- prove sources support claims;
- replace source-first checks;
- replace ethics, legal, journal, funder, client, or institutional review;
- guarantee assessment, publication, funding, or client acceptance.
After this skill, use academic-self-review-loop for formal drafts, then style and document-quality gates.