Skill Tracer
Skills › Scientific Agent Skills › research-grants › nstc_guidelines.md

nstc_guidelines.md

120 lines · skills/research-grants/references/nstc_guidelines.md at 5f2bc03

Open on GitHub
1# Taiwan NSTC proposal guidance
2
3Reviewed 2026-10-01. NSTC (國家科學及技術委員會, formerly MOST) requirements depend on
4the call, division, project type and year. Start with the
5[research-project area](https://www.nstc.gov.tw/folksonomy/list/3555dc27-f75a-491d-b9b2-ab66c156801f?l=ch)
6and the applicable call in the [solicitation directory](https://www.nstc.gov.tw/folksonomy/rfpList).
7The date convention matters: ROC year 115 is 2026 and 116 is 2027. Never turn an old
8114年度 deadline or August start date into a current application schedule.
9

Taiwan NSTC proposal guidance

Reviewed 2026-10-01. NSTC (國家科學及技術委員會, formerly MOST) requirements depend on the call, division, project type and year. Start with the research-project area and the applicable call in the solicitation directory. The date convention matters: ROC year 115 is 2026 and 116 is 2027. Never turn an old 114年度 deadline or August start date into a current application schedule.

Raw lines 1 to 9
10## Establish the applicable rules
121. Record call title, version/date, responsible division, project type, research field,
13 eligibility, application period and institutional deadline.
142. Retrieve the current [common forms](https://www.nstc.gov.tw/folksonomy/list/6c0b309f-518b-4613-a142-44af89109251?l=ch),
15 division-specific forms and special-call attachments. Record the downloaded filename
16 and revision so another reader can reproduce the requirements.
173. Read the [research-project operating rules, revised April 10, 2026](https://www.nstc.gov.tw/folksonomy/detail/2572b81b-2188-4fc5-af55-c2efac89c610).
18 Eligibility of the institution and PI, concurrent-project limits and special
19 program conditions need explicit checks; this guide is not an eligibility decision.
204. Have the institution reconcile call instructions with general rules. Do not use a
21 colleague's past successful proposal as the controlling form specification.

Establish the applicable rules

  1. Record call title, version/date, responsible division, project type, research field, eligibility, application period and institutional deadline.
  2. Retrieve the current common forms, division-specific forms and special-call attachments. Record the downloaded filename and revision so another reader can reproduce the requirements.
  3. Read the research-project operating rules, revised April 10, 2026. Eligibility of the institution and PI, concurrent-project limits and special program conditions need explicit checks; this guide is not an eligibility decision.
  4. Have the institution reconcile call instructions with general rules. Do not use a colleague's past successful proposal as the controlling form specification.
Raw lines 10 to 22
23## CM03 and associated forms
25CM03 is the research-content narrative. Abstracts and project information may belong
26in separate forms such as CM02; do not treat all application fields as CM03 sections.
27Use the exact current form headings and instructions. The
28[Humanities division](https://www.nstc.gov.tw/folksonomy/list/dea8c5cc-acd5-46d0-8dcb-15cfd8072e5d?l=ch)
29and [Engineering division](https://www.nstc.gov.tw/folksonomy/list/49ec3266-51af-4527-be73-fe5e10d0cf13?l=ch)
30publish separate materials. This skill does not assert a universal table of page
31limits, fonts, abstract lengths or keyword counts across divisions.
33Before writing, extract these exact constraints from the selected package:
35| Constraint | Record |
36| --- | --- |
37| Narrative | Required sections, language, page size, font, spacing and page maximum. |
38| Counted material | Whether references, figures, tables and attachments count toward limits. |
39| Abstracts | Chinese/English fields and their own character/word limits. |
40| Research period | Single/multi-year plan, each year's work and continuation requirements. |
41| Other forms | PI record, achievements, budget, ethics, collaboration and any special attachments. |
42| Upload | Permitted format, size, filename and system validation requirements. |
44A research-architecture diagram can clarify methods and dependencies, but is not a
45verified universal mandatory CM03 element. Community LaTeX/Overleaf classes are not
46NSTC-issued forms or proof of compliance. If one is used, compare the final output to
47the actual call template and visually check pagination and Chinese font rendering.
48No community template installation or compiled example is validated by this skill.

CM03 and associated forms

CM03 is the research-content narrative. Abstracts and project information may belong in separate forms such as CM02; do not treat all application fields as CM03 sections. Use the exact current form headings and instructions. The Humanities division and Engineering division publish separate materials. This skill does not assert a universal table of page limits, fonts, abstract lengths or keyword counts across divisions.

Before writing, extract these exact constraints from the selected package:

ConstraintRecord
NarrativeRequired sections, language, page size, font, spacing and page maximum.
Counted materialWhether references, figures, tables and attachments count toward limits.
AbstractsChinese/English fields and their own character/word limits.
Research periodSingle/multi-year plan, each year's work and continuation requirements.
Other formsPI record, achievements, budget, ethics, collaboration and any special attachments.
UploadPermitted format, size, filename and system validation requirements.

A research-architecture diagram can clarify methods and dependencies, but is not a verified universal mandatory CM03 element. Community LaTeX/Overleaf classes are not NSTC-issued forms or proof of compliance. If one is used, compare the final output to the actual call template and visually check pagination and Chinese font rendering. No community template installation or compiled example is validated by this skill.

Raw lines 23 to 49
50## Narrative development
52Adapt these writing prompts to the official headings; they are drafting advice.
54**Background and objectives (背景及目的)**
55- Define the problem, its importance, the precise knowledge gap and research question.
56- Critically compare relevant domestic and international work, including foundational
57 evidence as well as recent literature. Cite prior work and supervised theses accurately.
58- Explain what is new, why it is plausible, and how it advances the selected field.
59- For continuing work, distinguish completed results from new proposed activities.
61**Methods, steps and schedule (研究方法、進行步驟及執行進度)**
62- Connect each objective to a design, method, dataset/model, analysis and validation.
63- State assumptions, scientific uncertainties, feasible alternatives and decision points.
64- Explain required equipment/access and each year’s deliverable work.
65- Justify international collaboration or travel when requested in the application.
67**Expected work and outcomes (預期完成之工作項目及成果)**
68- Separate outputs (datasets, methods, papers, trained personnel) from longer-term
69 outcomes (uptake, policy or clinical benefit).
70- Give achievable metrics and a measurement plan. A promised paper count, journal rank,
71 licensing income or adoption target is not evidence it will happen.
72- Tie resources, personnel and timing to the proposed deliverables.
74A useful consistency chain is: problem -> objective -> methods -> work packages ->
75validation -> outputs. If one link is absent, revise the research plan before polishing
76language. Do not force a commercialization story onto basic research unless the call
77requires it. No universal passing score, second-project cutoff or success percentage
78is established here; use the program's published review rules and actual feedback.

Narrative development

Adapt these writing prompts to the official headings; they are drafting advice.

Background and objectives (背景及目的)

  • Define the problem, its importance, the precise knowledge gap and research question.
  • Critically compare relevant domestic and international work, including foundational evidence as well as recent literature. Cite prior work and supervised theses accurately.
  • Explain what is new, why it is plausible, and how it advances the selected field.
  • For continuing work, distinguish completed results from new proposed activities.

Methods, steps and schedule (研究方法、進行步驟及執行進度)

  • Connect each objective to a design, method, dataset/model, analysis and validation.
  • State assumptions, scientific uncertainties, feasible alternatives and decision points.
  • Explain required equipment/access and each year’s deliverable work.
  • Justify international collaboration or travel when requested in the application.

Expected work and outcomes (預期完成之工作項目及成果)

  • Separate outputs (datasets, methods, papers, trained personnel) from longer-term outcomes (uptake, policy or clinical benefit).
  • Give achievable metrics and a measurement plan. A promised paper count, journal rank, licensing income or adoption target is not evidence it will happen.
  • Tie resources, personnel and timing to the proposed deliverables.

A useful consistency chain is: problem -> objective -> methods -> work packages -> validation -> outputs. If one link is absent, revise the research plan before polishing language. Do not force a commercialization story onto basic research unless the call requires it. No universal passing score, second-project cutoff or success percentage is established here; use the program's published review rules and actual feedback.

Raw lines 50 to 79
80## Budget and ethics
82Under the April 2026 operating rules, research equipment means a directly relevant
83item costing **at least NT$10,000** with useful life **at least two years**. Use current
84institutional procurement and quotation requirements; NT$200,000 is not a universal
85price-comparison trigger established by this guide. Distinguish research personnel,
86consumables/other operating expenses, equipment, travel and management costs using
87the actual forms.
89The rules permit specified **research investigator fees (研究主持費)** for eligible
90funded projects. Do not equate these with ordinary salary, prohibit all PI payments,
91or assume multiple concurrent fees are allowed. Check the relevant project type,
92current rate, exclusions and institutional administration. Justify travel with actual
93purpose, itinerary and permitted rates rather than a canned NT$70,000-100,000 amount.
95The rules require first-time PIs and the specified first-time participating researchers
96to complete at least six hours of academic-ethics training within the three years
97before institutional submission and provide evidence to the institution. For newly
98hired first-time participating researchers after the project starts, check the
99three-month-from-hire requirement. Preserve the distinction between applicant and
100institutional certification responsibilities.
102Do not invent data, achievements, partner commitments, ethics approval or equipment
103access. The [NSTC academic-ethics rules](https://law.nstc.gov.tw/LawContent.aspx?id=FL044283)
104cover fabrication, falsification, plagiarism and inappropriate ghostwriting. Keep the
105investigator responsible for the substantive research plan and satisfy the institution's
106current authorship and AI-assistance rules.

Budget and ethics

Under the April 2026 operating rules, research equipment means a directly relevant item costing at least NT$10,000 with useful life at least two years. Use current institutional procurement and quotation requirements; NT$200,000 is not a universal price-comparison trigger established by this guide. Distinguish research personnel, consumables/other operating expenses, equipment, travel and management costs using the actual forms.

The rules permit specified research investigator fees (研究主持費) for eligible funded projects. Do not equate these with ordinary salary, prohibit all PI payments, or assume multiple concurrent fees are allowed. Check the relevant project type, current rate, exclusions and institutional administration. Justify travel with actual purpose, itinerary and permitted rates rather than a canned NT$70,000-100,000 amount.

The rules require first-time PIs and the specified first-time participating researchers to complete at least six hours of academic-ethics training within the three years before institutional submission and provide evidence to the institution. For newly hired first-time participating researchers after the project starts, check the three-month-from-hire requirement. Preserve the distinction between applicant and institutional certification responsibilities.

Do not invent data, achievements, partner commitments, ethics approval or equipment access. The NSTC academic-ethics rules cover fabrication, falsification, plagiarism and inappropriate ghostwriting. Keep the investigator responsible for the substantive research plan and satisfy the institution's current authorship and AI-assistance rules.

Raw lines 80 to 107
108## Submission and final checks
110Use the Academic Research Service Network (學術研發服務網) linked from NSTC. The PI
111prepares the application and sends it to the institution; the institution reviews and
112forwards it with required certifications. A PI's completed online draft does not prove
113NSTC receipt. Leave time for institutional review and confirm its earlier deadline,
114which can differ from the national deadline.
116Check that the selected call, year, division and research-field codes match the
117proposal; all required forms are included; budgets reconcile; cited evidence is real;
118Chinese and English summaries agree; and the exported files satisfy the actual limits.
119Save the submitted package and receipt only after authorized institutional submission.
120No authenticated portal, credentials, upload, or API workflow was tested in this refresh.

Submission and final checks

Use the Academic Research Service Network (學術研發服務網) linked from NSTC. The PI prepares the application and sends it to the institution; the institution reviews and forwards it with required certifications. A PI's completed online draft does not prove NSTC receipt. Leave time for institutional review and confirm its earlier deadline, which can differ from the national deadline.

Check that the selected call, year, division and research-field codes match the proposal; all required forms are included; budgets reconcile; cited evidence is real; Chinese and English summaries agree; and the exported files satisfy the actual limits. Save the submitted package and receipt only after authorized institutional submission. No authenticated portal, credentials, upload, or API workflow was tested in this refresh.

Raw lines 108 to 120