research-agent-starter-kit

Design and review active learning implementation support for university teachers, including classroom activity structures, preparation workflow, student participation, assessment alignment, and AI-agent support points.

Turn dissertation evidence into an AI agent design specification for supporting active learning implementation, including user scenarios, functions, interaction flow, boundaries, failure modes, and design rationale.

Synthesize confirmed co-design, design-elicitation, prototype-feedback, or participant-generated design outputs into requirements, design principles, prototype changes, participant contribution maps, and design decision logs.

Design specifications for dissertation figures such as conceptual frameworks, concept-card/interview process diagrams, AI-agent workflow diagrams, adoption-condition models, and findings maps.

Maintain a lightweight dissertation research wiki with project memory, literature notes, methodology decisions, supervisor feedback, design rationale, and open questions.

Audit an AI agent prototype or design concept for usability, pedagogical fit, teacher feedback, active learning support, evidence quality, and next design iterations.

Build and audit a teacher adoption conditions model for an AI agent supporting active learning, distinguishing perceptions, concerns, barriers, enablers, institutional conditions, and design implications.

Plan a teaching knowledge base for AI-agent support, including source types, metadata, RAG architecture, privacy boundaries, content governance, evaluation, and teacher-facing use cases.

Prepare viva or dissertation defense materials, likely examiner questions, concise answers, contribution statements, limitations, and presentation structure.

Decide which research project skills to use for each task, whether the task should be handled by one agent or split across multiple subagents, and how to coordinate outputs safely.

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.

Build and audit the controlling argument spine for this dissertation, proposal, literature review, methodology chapter, and major revisions so writing is motivation-led, source-grounded, and not shallow polishing.

Audit dissertation citations, bibliography entries, source metadata, quote accuracy, and whether cited sources actually support the attached claims.

Review formal research outputs before delivery, including proposal drafts, ethics/compliance materials, Word documents, source notes, reports, manuscripts, and stakeholder-facing summaries, using success criteria, evidence checks, formatting/render checks, and unresolved-field reporting.

Maintain compact-ready checkpoints, task state summaries, source maps, and handoff notes during long dissertation work so future turns preserve decisions, files changed, open questions, and evidence boundaries.

Use before presenting a formal Word, PDF, reviewer-facing, stakeholder-facing, or submission-facing research artifact as usable; checks pre-delivery lock evidence, final integrity/citation risks, skill receipts, DOCX structural parity, and layout regressions, with an explicit override path for acknowledged risk.

Use before delivering formal academic, reviewer-facing, stakeholder-facing, or submission-facing prose to scan and reduce overused binary negative-contrast constructions such as "rather than", "not...but", "不是...而是", and "而不是"; runs after self-review and before final delivery.

Use before substantive formal research drafting and again before delivery to check concrete integrity risks such as prompt residue, placeholders, fake or unverified references, unsupported claims, unresolved compliance requirements, and AI-use disclosure boundaries.

Use when a user asks to make research writing sound less AI-like, more human, more authorial, less generic, or asks about AI-writing, AIGC, detector scores, disclosure, de-AI, humanising, or lowering AI rate; improves academic voice and integrity without detector-evasion.

Use before delivering dissertation documents, proposal drafts, supervisor notes, or major chat answers when the user's preferred output style matters, especially to apply decision-first structure, two-track Thinking Pack/Decision Brief patterns, prohibited-phrase checks, and revision accountability.

Use when drafting, revising, or checking dissertation, thesis, manuscript, proposal, report, supervisor/PI/client/reviewer-facing, or academic/professional writing, with appropriate spelling, cautious stance, concise argument-led prose, and reduced generic AI-style phrasing.

Use before formal research writing or delivery to package source readiness, compliance or requirement evidence, citation boundaries, and unresolved confirmations for the artifact being moved forward.

Diagnose and recover from dissertation agent failures such as false runs, repeated tool loops, stale assumptions, context drift, wrong-window behavior, or environment-state mismatches before retrying.

Use before creating, copying, adapting, or updating project skills so new skills stay concise, non-overlapping, source-grounded, and compatible with the research agent rule stack; use with the system skill-creator when authoring SKILL.md files.

Use before claiming a GitHub, public template, or release-page update is complete, especially after version bumps, tags, GitHub Releases, README badges, About/sidebar text, topics, links, or public onboarding changes.

Use when creating, revising, or auditing high-impact journal style scientific figures, Nature-style figure logic, multi-panel research figures, manuscript-ready plots, figure contracts, panel narratives, SVG/PDF/TIFF export checks, or publication-grade visual argumentation.

Use when planning, drafting, restructuring, or polishing high-impact journal style academic prose, Nature-style manuscript sections, article abstracts, introductions, discussions, contribution statements, or publication-leaning research writing.

Use when creating, planning, auditing, or choosing tools for neural-network architecture figures, CNN diagrams, AI model schematic diagrams, NN-SVG, PlotNeuralNet, draw_convnet, LaTeX/TikZ neural-network figures, or publication-ready model architecture visuals.

Use before delivering proposal, literature review, methodology, manuscript, report, grant, supervisor-facing, reviewer-facing, client-facing, or formal academic/professional drafts to run a two-pass self-review and revision loop using intrinsic writing-quality criteria before style and document-quality gates.

Use for structured research project or agent-system ideation when the user's idea is still unclear, high-impact, or needs route comparison before drafting, implementation, or skill changes.

Audit the dissertation agent's rule stack, memory layers, skill routing, tool discipline, window separation, persistence, and output gates for conflicts, stale context, hidden assumptions, or maintenance risk.

Evidence-first audit of dissertation automations, hooks, scheduled checks, MCP/connectors, browser/LMS monitors, and workflow wrappers to identify what is live, broken, redundant, missing, or unsafe before enabling anything.

Create or revise dissertation chapter outlines, argument maps, section plans, and writing schedules for an education research dissertation on AI agents and active learning.

Manage dissertation project knowledge across research-wiki, knowledge-base, Obsidian, source registers, LMS notes, supervisor or public resource notes, and task-state files with deduplication, source-of-record rules, indexes, and privacy boundaries.

Use when dissertation work should convert new reading, source searches, LMS/module materials, supervisor resources, or discussion outcomes into durable project knowledge, literature-map updates, proposal implications, and next-reading questions without enabling unsupervised background automation.

Plan, review, and synthesize literature for a dissertation on AI agents, active learning, higher education, teacher adoption, co-design, educational technology, and responsible AI in teaching.

Critically review dissertation research design, research questions, theoretical framing, methodology, findings, discussion, or chapter drafts for an education dissertation on AI agents, active learning, teacher perceptions, concerns, and adoption conditions.

Plan, run, and record source-grounded literature or web research for the dissertation, with search questions, source screening, evidence boundaries, citation readiness, and knowledge-base updates.

Shared dissertation protocols for privacy, citation discipline, review independence, evidence handling, output versioning, and responsible use of AI assistance in an education research dissertation.

Audit dissertation project skills for trigger clarity, overlap, stale content, missing safety gates, maintenance value, and merge/keep/improve/retire decisions.