Hello, this is Ryuta Hamamoto from TIMEWELL.
If you have just been handed research security at one of Japan's national research and development agencies (national R&D agencies), the information systems team may well tell you, "We already have a policy that follows the government's Common Standards." Fair enough. So what does research security add on top of that? And does it line up with how export control treats the cloud and laptops taken abroad? That is usually where the conversation stalls.
My answer: before drafting any new rule, run the "verification" that note 33 of the Cabinet Office research security procedures manual asks for, and keep it as a record for each research project. A national R&D agency is an incorporated administrative agency, so for you that verification sits on the "necessary" side. What the note asks is narrower than people assume, though. You check whether information on critical technologies handled in a Specified Research and Development Program is secure against your own information security policy and operating rules. Nobody in the research security office has to work through the Common Standards from scratch.
The short version
- Note 33 is attached to "strengthening measures against cyberattacks," one of the example risk mitigation measures in Part II, Chapter 3, section 3-4 of the manual (p. 21). For "agencies" (national administrative organs, incorporated administrative agencies and designated corporations), verification is "necessary." For other research institutions, it "may be considered."
- National R&D agencies count as "agencies" because Article 2(1) of the Act on General Rules for Incorporated Administrative Agencies makes them a type of incorporated administrative agency, and the Basic Act on Cybersecurity uses that definition when it gives the Cybersecurity Strategic Headquarters the job of setting and evaluating (including auditing) cybersecurity standards for those agencies.
- For research data, exchanges with partners, cloud, web meetings, outsourcing, equipment purchases and taking data abroad, research security (the manual), export control (FEFTA Article 25) and information security (your own rules) each ask a different question. One table that puts the three side by side, plus one verification record per project, keeps things from falling through the gaps.
One premise first. The manual states that, in judging whether a counterpart is a trustworthy partner, "it is a matter of course that there must be no discriminatory treatment on grounds of nationality, race, religion, culture or the like" (Part I, Chapter 1, 1-1(2), p. 3).1 The work in this article is not about drawing lines by who people are. It is about checking, against your rules, which information sits where and who can use it. Being on a country's restricted list is a regulatory designation, not a judgment on the listed organisation or person.
I have covered the seven example measures and the different weight of the Common Standards and TICS in implementing research security risk mitigation, and the article-by-article map for merging export control and research security rules in one rulebook for national R&D agencies. This piece deals only with the seam between those two: where note 33 meets the information security rulebook.
What note 33 asks for, read word by word
Start with where the note sits. Section 3-4 ("risk mitigation measures") of Part II, Chapter 3 (measures by research institutions) says institutions must take risk mitigation measures in light of their risk identification and assessment. It then says the measures "may be those exemplified below, and reasonable measures commensurate with the degree of risk are sufficient," and lists seven examples. The seventh is "strengthening measures against cyberattacks," and note 33 hangs off it (pp. 20-21).1 The note's heading is "risk management relating to information security." Here is the text in my translation:
"For national administrative organs, incorporated administrative agencies and designated corporations (hereinafter 'agencies'), the 'Common Standards' have been prepared so that the information security measures commonly required of all agencies can be taken. Accordingly, agencies must verify, against the information security policies, operating rules and the like that each agency has set based on the Common Standards, whether information security is ensured for information on critical technologies handled in a Specified Research and Development Program. Research institutions (other than agencies) may also consider verifying whether information security is ensured, with reference to the Common Standards. It may also be considered to refer to the Common Standards for the following typical cases expected at the research and development stage." The note then lists three cases (outsourcing the development, operation or maintenance of information systems or applications; using cloud services such as virtual servers and web meeting services; procuring IT equipment), and closes with a reference for higher education institutions: the National Institute of Informatics' sample rules for information security at higher education institutions.
The manual defines its own wording in Part I, Chapter 1, 1-2(2). Where it says "necessary," it describes a minimum measure; where it says "desirable," it describes a measure that would be good to take but is not required for now (p. 4).1 "May be considered" does not appear in that definition. Split the note sentence by sentence and you get this:
| Sentence | Subject | What | Wording | Where it sits under Part I 1-2(2) |
|---|---|---|---|---|
| 1st | Agencies | The Common Standards exist | "have been prepared" | A statement of fact, not a measure |
| 2nd | Agencies | Verify, against your own policies and operating rules, that information on critical technologies is secure | "necessary" | Minimum measure |
| 3rd | Research institutions other than agencies | Verify with reference to the Common Standards | "may be considered" | Neither of the two defined words |
| 4th | No subject | Refer to the Common Standards in the three cases | "may be considered" | Same |
| 5th | Higher education institutions | Also refer to the NII sample rules | "may be considered" | Same |
I read "may be considered" as wording that offers examples or points to a reference. That is my reading; the manual does not define it.
The second sentence carries three conditions that summaries tend to drop. First, scope: the target is "information on critical technologies handled in a Specified Research and Development Program," not everything the agency holds. Second, the yardstick: you check against "the information security policies, operating rules and the like that each agency has set based on the Common Standards," not against the Common Standards directly. Third, the act: the note asks you to "verify whether information security is ensured." It does not ask for new rules. Drop those three and you end up with a much heavier reading than the text supports, something like "a national R&D agency must write new rules aligned with the Common Standards for research security."
Note 33 is attached to an example measure, but the second sentence does not say "only if you choose cyberattack countermeasures as your measure." At a national R&D agency I would run the check whichever measures you choose. That is my recommendation, not the manual's words. The fourth sentence, on the three cases, has no subject and uses "may be considered," so I read direct reference to the Common Standards as not "necessary" even for agencies. In practice, though, if your own policy already covers outsourcing or the cloud, you will check against those provisions as part of the second-sentence verification anyway.
You can see the split was deliberate by comparing the draft. The draft manual submitted to the expert panel's fifth meeting (18 July 2025) said that agencies and other research institutions alike were "required to verify" (draft p. 25, a figure quoting p. 10 of the fourth meeting's materials).2 The final version of December 2025 split that into "necessary" for agencies and "may be considered" for everyone else.
How will funding agencies ask about this in their questionnaires? Note 15 of the manual says the questionnaire checks whether institutions are implementing "the minimum measures (the measures described as 'necessary' in this manual)" (Part I, Chapter 2, 2-5(3), p. 9). Yet the questionnaire template in Annex 3 has no question on the note 33 check itself. "Strengthening measures against cyberattacks" appears only as one of the tick-box options for risk mitigation measures in 4(1) (p. 43). The Annex 2 checklist for lead institutions mentions it only as an example in the footnote to item 3 (p. 31).1 Funding agencies are told to "modify the template as appropriate to the content of the program" (p. 39),1 so whether a real questionnaire asks about it has to be checked call by call. Even if nobody asks, I think you should keep a record that you did the check.
Note 33 may also grow. The Science, Technology and Innovation Basic Plan, adopted by the Cabinet on 27 March 2026, says in Chapter 4, section 3 ("strengthening research security") that, to prevent leaks of research data in Specified Research and Development Programs, "relevant ministries will work together to consider the necessary additions to the procedures manual so that universities, national R&D agencies and others can strictly handle research data in line with the nature of each program, based on the approach of the 'Common Standards for Cybersecurity Measures for Government Agencies and Related Organisations' (Cybersecurity Strategic Headquarters, 27 June 2025)" (pp. 48-49, my translation).3 The Cabinet Office's explanatory slides on the manual quote the same sentence on a page headed "(reference) strengthening cybersecurity" (p. 13).4 The copy of the manual I downloaded again from the Cabinet Office site on 22 November 2026 was the same file as the December 2025 version. When it changes, I will update the tables here.
Why national R&D agencies fall under the Common Standards
Note 33 names "national administrative organs, incorporated administrative agencies and designated corporations." It never says "national R&D agency." You get there by linking two statutes, and that is also the order I would use when explaining it to your information systems team or your auditor.
Article 2(1) of the Act on General Rules for Incorporated Administrative Agencies defines an incorporated administrative agency as a body "established as a medium-term objective management agency, a national research and development agency or an administrative execution agency, as provided by this Act and the individual Act." Article 2(3) defines the national research and development agency as an incorporated administrative agency, set up by its individual Act, whose main business is research and development and whose purpose is to secure the maximum results from it.5 A national R&D agency is one of the three types.
Article 13 of the Basic Act on Cybersecurity (ensuring cybersecurity at national administrative organs and the like) defines incorporated administrative agencies by reference to "Article 2(1) of the Act on General Rules for Incorporated Administrative Agencies" and says the State will take measures including "the formulation of uniform standards on cybersecurity" for national administrative organs, incorporated administrative agencies and designated corporations. Designated corporations are special corporations and authorized corporations that the Cybersecurity Strategic Headquarters designates as needing further State measures, having regard to the impact on people's lives or economic activity if their cybersecurity were not ensured (same Article).6 Note 33's "agencies" are exactly that trio.
Article 25(1) of the same Act lists the Headquarters' duties. Item (ii) covers "matters relating to the preparation of standards for cybersecurity measures at national administrative organs, incorporated administrative agencies and designated corporations, the evaluation (including audits) of measures based on those standards, and other promotion of measures based on those standards." Item (iv) covers evaluating the state of cybersecurity at the same three groups, including monitoring and analysis of unauthorized activity and evaluation of measures taken on serious incidents (including investigations into causes). Article 31(1)(i) lets the Headquarters entrust part of that work, including audits under item (ii) and investigations under item (iv) that relate to incorporated administrative agencies and designated corporations, to the Information-technology Promotion Agency and other bodies named by Cabinet Order. Under Article 33(1), the Headquarters may ask the heads of incorporated administrative agencies and others for documents, explanations and other cooperation, and those asked "must comply unless there is a justifiable reason."6 The version I checked is the one in force from 1 October 2026, which reflects the amendments made by Act No. 43 of 2025.
The same Act shows how universities differ. Article 8 asks "universities and other educational and research institutions" to endeavor, on their own initiative, to ensure cybersecurity. Article 33(1) lists the presidents of national university corporations and the heads of inter-university research institute corporations separately from the heads of incorporated administrative agencies.6 So when a national R&D agency and a university work on the same data in a joint project, the weight of note 33 can differ between them. Whether a national university corporation counts as an "agency" depends on whether it has been designated; I have not checked the list of designated corporations.
What, then, do the Common Standards themselves say? The Common Standards page of the National Cyber Office (formerly NISC) states that the Cybersecurity Strategic Headquarters adopted the Common Standards on 27 June 2025 under Article 25(1)(ii) of the Basic Act on Cybersecurity. It describes them as a common framework for raising the information security level of "national administrative organs and incorporated administrative agencies, etc." The latest documents are the Uniform Norms, the Common Standards (FY2025 edition) and the Guidelines for Formulating Countermeasure Standards (FY2025 edition), and the Guidelines were partly revised in September 2025 and June 2026.7 The adoption date matches the Basic Plan.3 I have not, however, read the documents themselves (the PDFs). For that reason, this article gives no article or section numbers from the Common Standards and no names of the roles they define.
What matters in practice is that note 33's yardstick is your own policy. Your agency should already have an information security policy and operating rules built on the Common Standards, and note 33 asks you to check against those. The research security office's job is to decide what is being checked (which project, which information), ask the information security team to check it, and keep the result on file. Reading the Common Standards article by article is work the policy owners already do.
Replace siloed classification work with AI.
METI's FY2024 data shows 52% of foreign exchange law violations stem from classification errors. Download the TRAFEED product catalog covering features and rollout.
Scenario table: research data, partners, cloud and travel
The same act raises three different questions. The research security manual asks whether your measures are reasonable for the level of risk. FEFTA Article 25 asks whether you are carrying out a transaction to provide controlled technology ("specified technology") in a foreign country or to a non-resident, or exporting a record of it or sending it abroad, and if so whether you hold a license. Your information security policy asks whether the information is protected as the rules say. Try to answer all three with one answer and something slips.
Some FEFTA vocabulary first. Article 25(1) requires a license from the Minister of Economy, Trade and Industry for "transactions intended to provide specified technology in a specified foreign country" and for "transactions intended to provide specified technology to a non-resident of a specified foreign country." Article 25(3) allows the Minister, where needed to ensure that paragraphs (1) and (2) are reliably implemented, to impose a license requirement by Cabinet Order on exporting documents, drawings or recording media containing specified technology, and on sending it by telecommunications for receipt abroad (limited to transmissions from telecommunications facilities in Japan).8 METI's Services Circular defines "provide" as "placing it in a state where others can use it" (1(3)(ta), p. 3),9 and METI's fifth-edition guidance for universities and research institutions lists "experimental data" among its examples of technical data (III 3(2), p. 38).10 Depending on the content, research data can be controlled technology.
| Scenario | Research security (manual) | Export control (FEFTA, Services Circular, fifth edition) | Information security (note 33 and your own rules) | Main owners (my proposal) |
|---|---|---|---|---|
| Storing research data and access rights | Collecting information on data management plans (DMPs), information systems and facilities as a routine is "desirable" (II 3-1(3)). Checking the project DMP and three management policies at application is also "desirable" (3-3(1)(ii)). Managing access rights is an example measure in 3-4 | Experimental data is an example of technical data (fifth edition p. 38). Access to list-controlled technology information should be managed (p. 72). Placing it where others can use it counts as "providing" (Services Circular 1(3)(ta)) | Verification of information on critical technologies against your policies and rules is "necessary" (note 33, 2nd sentence) | Research promotion (identifying the target), information security (verification), export control (classification) |
| Exchanging data with partners (email, shared folders) | Checking that joint research and other agreements are appropriate on matters including "access to research data" is "necessary" (3-3(1)(v); partner institutions under (2)(iii)) | Email is one of the listed occasions for providing technology (fifth edition p. 34). Check whether a license under Article 25(1) or (3) is needed: is the recipient a non-resident, is it provided abroad, is it sent for receipt abroad? | Check against your rules on external exchanges. I have not verified the matching parts of the Common Standards | Industry collaboration and legal (contracts), export control, information security |
| Cloud (virtual servers, storage) | Checking that the management policy for the information systems used in the research (including those the PI manages) is appropriate is "desirable" (3-3(1)(ii)(3)) | A contract only for storing information for your own use is, in principle, not a services transaction even if the servers are abroad. If you sign knowing the provider can view, obtain or use the technology, it is treated as a transaction to provide it (Services Circular, Attachment 1-2(1)) | One of the typical cases where referring to the Common Standards "may be considered" (note 33, 4th sentence) | Information security (contract and configuration), export control (access clauses) |
| Web meetings | Considering who attends meetings in light of the sensitivity of the information is an example measure in 3-4 | Providing technology in meetings counts whether in person or online (fifth edition p. 34). Joining a closed meeting abroad online and providing controlled technology needs an Article 25(1) license, just as if you had travelled (p. 82) | Note 33 names web meeting services as an example of cloud services | PI (who attends), export control, information security (the service) |
| Outsourcing development, operation or maintenance of systems | Where a project member is an employee of a contractor or similar, checking the contract for their participation is "desirable" (note 32 to 3-4) | If the contractor is placed where it can use specified technology, "providing" becomes an issue (my reading of Services Circular 1(3)(ta)) | One of the typical cases where referring to the Common Standards "may be considered" (note 33, 4th sentence) | Information security (vendor review), export control |
| Procuring IT equipment | Nothing beyond note 33 | Buying equipment is not in itself a transaction to provide technology. If the vendor connects remotely for maintenance, treat it like the outsourcing row (my reading) | One of the typical cases where referring to the Common Standards "may be considered" (note 33, 4th sentence) | Information security, procurement |
| Business travel and taking laptops or USB drives abroad | The manual says nothing on this | Taking a USB drive or papers containing controlled technology abroad, or sending it abroad, in order to provide it to someone there needs an Article 25(3) license unless you already hold an Article 25(1) license or a similar exception applies, and you still need an Article 25(1) license before actually providing it (fifth edition p. 82). Off-the-shelf laptops are, with some exceptions, not list-controlled. The guidance asks you to encrypt or password-protect laptops and USB drives holding controlled technology data (p. 83) | Check against your rules on taking devices and media off site. I have not verified the matching parts of the Common Standards | Export control (pre-travel check), information security, PI |
The manual column refers to items in Part II, Chapter 3 (including 3-4 and note 32 on pp. 20-21).1 The two cells marked "my reading" are my application of the Services Circular's definition, not something METI has written.
Cloud is where the answers diverge most. Attachment 1-2(1) of the Services Circular notes that with storage services, information may end up on servers abroad whether or not the user intends it, and then says that a contract whose only purpose is storage for your own use is, in principle, not a services transaction under FEFTA Article 25(1) and not caught by Article 25(3) either. Three qualifications follow. First, if the contract is in substance a transaction to provide specified technology to the provider, it is a services transaction; for example, if you sign knowing that the provider can view, obtain or use the stored technology, the contract is treated as one to provide it. Second, if after signing you find that the provider is viewing, obtaining or using the stored technology and you keep the contract going, the transaction is treated as starting once the time needed to delete that technology has passed. Third, if you use storage to provide technology to a third party, that is naturally a transaction to provide it to that third party (p. 7).9 Putting data in the cloud to share with a project partner falls under the third.
Even when export control's answer is "no license, in principle," the note 33 check still has to happen. FEFTA looks at whether you are handing technology to someone; information security looks at whether the way you store it follows your rules. The research security manual sits between them, calling it "desirable" to check the management policy for information systems at application. One contract, three teams, three yardsticks. Build that into the procedure from the start and no team's "it's fine" quietly stands in for the other two answers.
Web meetings are where all three apply together. The manual lists considering who attends as an example measure, export control treats explaining technology in a meeting as providing it whether in person or online, and note 33 names web meeting services as a case for referring to the Common Standards. Taking data abroad is the opposite: the manual says nothing about it. Export control and information security both still apply, so put both into the pre-travel check and fewer things slip.
The contract row is easiest to fill with a form. The joint and contract research checklist with model clauses (Excel) has 10 pre-contract checks, a screening record for the partner organisation, seven model clauses, and fields for recording samples, equipment and data you send. When you draft the clause on access to research data, you can put it in the same field as the export control check.
Who owns what: research promotion, export control, information security
The manual says that institutions whose researchers apply, or plan to apply, to a Specified Research and Development Program need to set up a structure for managing the information required for risk management, and that a structure where relevant departments collect the information together and manage it centrally is "desirable" (II 3-1(2), p. 17).1 On which department should lead, the Cabinet Office Q&A answers that "relevant departments already handling security export control, research integrity and the like would likely be at the center," and that if no suitable department exists, one capable of the work must be designated, including through reorganisation (Part II No. 9).11
Notice that the answer does not mention the information security department. Research security tends to settle in the export control or research integrity office, with information security treated as "someone to call if something comes up." At a national R&D agency, though, the note 33 check is on the "necessary" side, and its yardstick is the policy and rules that information security owns. I think a national R&D agency should write the information security team into the research security procedure as a fixed step. Leave them as a contact point and no one can later show whether the check was done.
Export control leads to the same place. In the information management item under "other procedures" for checking technology transfers, the fifth-edition guidance says "list-controlled technology information in particular should be access-managed," and that where information security rules already exist, "applying them mutatis mutandis to manage access to technical information is effective" (IV 2(4), p. 72).10 Having the export control rules hand access management over to the information security rules is the approach the guidance itself describes. The Basic Plan adds that the government will support universities, inter-university research institutes and national R&D agencies with advice and information sharing on cybersecurity measures and structures, training and exercises, and advice when incidents occur, and that research institutions running Specified Research and Development Programs "need to implement appropriate cybersecurity measures, and the ministries responsible for the programs will respond appropriately" (p. 48).3
Here is one way to split the work. The allocation is my proposal; neither the manual nor the fifth edition sets it.
| Task | Research promotion (research security) | Export control | Information security | PI |
|---|---|---|---|---|
| Identify critical-technology information per project | Coordinate | Supply classification results | Reflect in the agency's information categories | Define the scope of managed data |
| Note 33 check | Request and record | Not involved | Check against policy and operating rules | Declare the information systems used |
| Review of cloud, web meetings and outsourcing | Not involved | Access clauses and whether technology is provided | Review contract and configuration | Apply to use |
| Data clauses in joint research agreements | Check the four contract items | Whether technology is provided, and licenses | Check the means of sharing | Decide what is shared |
| Taking data abroad before a trip | Not involved | Whether a license is needed | Check device encryption and similar | Declare what is being taken |
| A leak or suspected leak | Report to parties including the funding agency | Whether a report is needed under the export control rules | First response | First notice |
"Define the scope of managed data" in the PI column follows note 34 of the manual, which describes managed data as "research data whose scope the researcher defines as subject to management and use, based on the standards of the institution conducting research and development, such as a university, inter-university research institute corporation or national R&D agency, or of the funding agency" (p. 22).1 The reporting row follows the Q&A, which says that whether or not the manual applies, "reporting as quickly as possible to parties including the funding agency is necessary" to prevent damage from spreading (Part I No. 7).11
National R&D agencies that also fund research have one more contact point. The manual requires funding agencies to set up their operating structure before a call opens and lists, among the points to observe, "formulating a policy for managing information, research data and the like obtained from the receipt of applications onward" (Part II, Chapter 2, 2-1(1)(3), p. 13).1 That means deciding which information systems hold other institutions' information received in review, and who can see it, and it needs the information security team too. I explained why to keep that separate from the research promotion ledger in the section "Why the funding-agency ledger stays separate" of one rulebook for national R&D agencies.
To put this split on a calendar, use the national R&D agency research security 90-day plan (Excel). It lays out the nine extra procedures the manual adds (including data management) with owner, deadline and evidence fields, and has a sheet for a five-step emergency contact flow. Add the information security team in the owner field and the split above goes straight onto the schedule.
Five steps for checking your existing rules
Start the review by connecting the provisions you already have, before writing new ones. I suggest five steps; they are my proposal.
Step one: decide the target information for each project. Note 33 covers "information on critical technologies handled in a Specified Research and Development Program," so for each project you plan to submit, write down which data falls in that scope. Start from the managed data in the data management plan (DMP) that the PI defines, and attach export control's classification result. Then research security and export control stop calling the same data by different names.
Step two: go through your information security policy and operating rules and pick out the provisions that match the seven scenarios in the table above: storage and access rights, external exchanges, cloud, web meetings, outsourcing, procurement and taking data abroad. Note the article number and the version of the rules. Where you cannot find a provision, leave the cell blank on purpose. A blank is exactly where you cannot yet answer note 33's question of whether security is ensured.
Step three: check against the export control rules. Does the access management provision for technical information apply the information security rules (the approach on p. 72 of the fifth edition)? Does cloud contract review include checking whether the provider can view, obtain or use the data (Services Circular, Attachment 1-2)? Does the pre-travel check cover whether taking or sending data abroad needs a license (fifth edition p. 82)? Those three points.
Step four: write a "verification step" into the research security procedure. Put the data management check at application (II 3-3(1)(ii), desirable) and the note 33 check (necessary) on one record. If you tick "strengthening measures against cyberattacks" in 4(1) of a funding agency's questionnaire, keep the basis for it on the same record. Step five: set the record and the review dates. Keep the date, who checked, the name and version of the rules checked against, the result, and any gaps with the action taken. Review it when you add project members, when the manual is revised and when your own policy changes.
A line of the verification record might look like this. Every entry is invented.
| Field | Example (fictional) |
|---|---|
| Project | Project A, applying to a Specified Research and Development Program (the record carries the full title) |
| Target information | The part of the experimental data the PI defined as managed data. Classified by export control; result attached |
| Rules checked against | Information security policy and its operating rules (record the version and revision date of each) |
| Result by scenario | Storage: secure. Cloud: access clauses in the contract checked. Web meetings: PI has set who attends. Travel: no trips planned |
| Gaps and actions | The shared folder with the partner has broad permissions; narrow it before the project starts |
| Checked by and date | Names of the information security and research promotion staff, and the date |
| Next review | When project members are added, or when the manual or the policy is revised |
The manual leaves the placement of rules to each institution. My recommendation is not to write any control twice. Keep access control, outsourcing, cloud, procurement and travel controls in the information security policy and operating rules, which are note 33's yardstick. Put an article in the export control rules applying the information security rules to access management for specified technology. In the research security chapter, write only who carries out the note 33 check, when, and how it is recorded. Write the same control into three rulebooks and the other two drift as soon as one is revised.
Model text (an article for the research security chapter; my draft, to be reviewed by your legal team) Article X (Information security verification) The research promotion department shall request the information security department to verify, against the [name] information security policy and the related operating rules, whether information security is ensured for information on critical technologies handled in a Specified Research and Development Program. 2. The information security department shall record the result of the verification in the preceding paragraph for each research project and notify the research promotion department. 3. The verification in the preceding two paragraphs shall be carried out before applying to a Specified Research and Development Program, when project members are added, and when the rules referred to in paragraph 1 are revised.
If you want the questionnaire draft and the Annex 2 self-check in one ledger, use the research security self-check and due diligence record kit (Excel). It sets out the questions from the three Annex 2 checklists in the original wording, sends any row marked "partly" or "cannot" to a planning sheet, and includes a sheet for drafting questionnaire answers. The verification record above can sit behind those answers as the evidence.
Summary: what to do in the first two weeks
For a national R&D agency, the note 33 check is a minimum measure. Even so, you do not have to start by writing rules. In the first two weeks, do three things. List the projects you plan to submit and write down the critical-technology information in each. Sit down with the information security team, pull the provisions for the seven scenarios out of your policy and operating rules, and find the blanks. With export control, confirm that cloud access clauses and pre-travel checks on taking data abroad are actually in the rules.
The three regimes ask different questions about the same data. Do not let one team's "no problem" stand in for the other two. The device for that, as I see it, is one verification record per project. With the Basic Plan saying the ministries will consider adding to the manual, note 33 may become more specific. Keep a record of today's check so that you do not have to rebuild everything when it does.
What I have not verified
As of 22 November 2026, I have not verified the following: the text of the government's Common Standards (the uniform norms, the uniform standards and the guidelines for setting measures standards), their article and section numbers, the roles they define, and what they say on outsourcing, cloud and taking data off site; the list of designated corporations, and whether national university corporations count as "agencies"; how individual national R&D agencies structure their information security policies; the timing and content of the "necessary additions to the procedures manual" mentioned in the Basic Plan; and whether the two cells marked "my reading" in the table match METI's interpretation. Once I can check the Common Standards in the original, I will add the matching provisions to the information security column of the table.
Read next
- Implementing research security risk mitigation: the seven example measures and the different weight of the Common Standards and TICS
- One rulebook for national R&D agencies: article mapping, the statement of operational procedures and approval
- Building a research security system: the extra steps for national R&D agencies, a 90-day schedule and an emergency contact flow
- Ten checks before a joint or contract research agreement with a foreign institution: clauses on access to research data alongside the export control check
- Self-checking with the Annex 2 checklist before you apply: the Annex 2 questions one by one
If you are looking for a way to keep per-project verification records and screening records for partners and related parties in one ledger, you can contact us about TRAFEED.
Footnotes
-
Expert Panel on Ensuring Research Security and Research Integrity, "Procedures Manual for Ensuring Research Security," December 2025 (Cabinet Office, in Japanese). Part I, Chapter 1, 1-1(2) (no discriminatory treatment, p. 3) and 1-2(2) (definitions of "necessary" and "desirable," p. 4); Part I, Chapter 2, 2-5(3) and note 15 (measures checked by the questionnaire, p. 9); Part II, Chapter 2, 2-1(1)(3) (funding agencies' policy for managing information and research data, p. 13); Part II, Chapter 3, 3-1(2)(3) (p. 17), 3-3(1)(ii)(v) and (2)(iii) (pp. 19-20), 3-4 with notes 32 and 33 (pp. 20-21), note 34 (managed data, p. 22); Annex 2 checklist for lead institutions, item 3 and its footnote 3 (p. 31); Annex 3, instructions for use (p. 39) and 4(1) (p. 43). Downloaded again from the Cabinet Office site on 22 November 2026 and confirmed to be the same file as my copy. https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/guidelines_v1.pdf ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9
-
Expert Panel on Ensuring Research Security and Research Integrity, fifth meeting (18 July 2025), material "Procedures Manual for Ensuring Research Security (draft): Guidelines for Preventing Leaks of Critical Technologies," p. 25 (a figure quoting p. 10 of the fourth meeting's materials) (in Japanese). Checked against my copy (retrieved 21 November 2026, compared 22 November 2026). https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/5kai/shiryo1.pdf ↩
-
"Science, Technology and Innovation Basic Plan," Cabinet decision of 27 March 2026, Chapter 4, section 3 ("strengthening research security," pp. 47-49) (in Japanese). Checked against my copy (retrieved 21 November 2026, compared 22 November 2026). https://www8.cao.go.jp/cstp/kihonkeikaku/7honbun.pdf ↩ ↩2 ↩3
-
Cabinet Office, Science, Technology and Innovation Promotion Secretariat, explanatory slides "On the Procedures Manual for Ensuring Research Security," p. 13, "(reference) strengthening cybersecurity" (in Japanese). Retrieved 22 November 2026. https://www8.cao.go.jp/cstp/kokusaiteki/integrity/explanatory_guidelines.pdf ↩
-
e-Gov, "Act on General Rules for Incorporated Administrative Agencies" (Act No. 103 of 1999), Article 2(1) and (3) (in Japanese). Version in force from 1 June 2025, retrieved through the e-Gov Laws API (v2) on 22 November 2026. https://laws.e-gov.go.jp/law/411AC0000000103 ↩
-
e-Gov, "Basic Act on Cybersecurity" (Act No. 104 of 2014), Articles 8, 13, 25(1)(ii) and (iv), 31(1) and 33(1) (in Japanese). Version in force from 1 October 2026, reflecting the amendments by Act No. 43 of 2025, retrieved through the e-Gov Laws API (v2) on 22 November 2026. https://laws.e-gov.go.jp/law/426AC1000000104 ↩ ↩2 ↩3
-
National Cyber Office, "Measures for Government Agencies," page on the Common Standards for Cybersecurity Measures for Government Agencies and Related Agencies (in Japanese). The adoption date and legal basis, the latest documents, and the partial revisions of the Guidelines (September 2025 and June 2026) were checked on the page itself on 22 November 2026; the PDFs of the documents were not read. https://www.cyber.go.jp/policy/group/general/kijun.html ↩
-
e-Gov, "Foreign Exchange and Foreign Trade Act" (Act No. 228 of 1949), Article 25(1) and (3) (in Japanese). Version in force from 12 August 2026, retrieved through the e-Gov Laws API (v2) on 22 November 2026. https://laws.e-gov.go.jp/law/324AC0000000228 ↩
-
METI, Services Circular (the circular on transactions or acts providing technology that require permission under FEFTA Article 25(1) and Article 17(2) to (4) of the Foreign Exchange Order, 4-Bōkyoku No. 492; last amended by Export Notice 2025 No. 27, promulgated 14 November 2025, effective 14 February 2026) (in Japanese). 1(3)(ta) (definition of "provide," p. 3) and Attachment 1-2(1) (interpretation for storage services, p. 7). Checked against my copy (retrieved 20 November 2026, compared 22 November 2026). https://www.meti.go.jp/policy/anpo/law_document/tutatu/t10kaisei/ekimu_tutatu.pdf ↩ ↩2
-
METI Trade Control Department, "Guidance on Sensitive Technology Management for Security Trade Control (for Universities and Research Institutions)," fifth edition, September 2025 (in Japanese). Printed pages cited: p. 34 (opening of III: occasions for providing technology, in person and online), p. 38 (III 3(2), forms of provision and examples of technical data), p. 72 (IV 2(4), information management), pp. 82-83 (V 2, overseas travel and temporary return home). Checked against my copy (retrieved 20 November 2026, compared 22 November 2026). https://www.meti.go.jp/policy/anpo/daigaku/guidance5.pdf ↩ ↩2
-
Cabinet Office, "Q&A on the Procedures Manual for Ensuring Research Security," last updated 26 August 2026 (in Japanese). Part I No. 7 (reporting a leak), Part II No. 9 (which department handles technology leak prevention). Checked against my copy (22 November 2026). https://www8.cao.go.jp/cstp/kokusaiteki/integrity/security_qa_v1.pdf ↩ ↩2






