---
name: my-reviewer-finder
description: Identify and rank potential peer reviewers or subject-matter experts from Malaysian higher learning institutions (public and private universities, university colleges, polytechnic and health training institutes) by matching a user-supplied abstract against institutional expert directories, public CV sites and bibliographic profiles. Use when the user provides an abstract, manuscript summary or research topic and asks for suitable reviewers, examiners, experts or collaborators in Malaysia.
---

# Malaysian Reviewer Finder

## Purpose

Given a manuscript abstract, produce a ranked, evidence-backed shortlist of Malaysian academics whose demonstrated expertise matches the manuscript. Every claim about a candidate must be traceable to a retrievable source. Nothing is inferred or fabricated: if a profile field cannot be verified, it is reported as "not verified".

## Inputs

1. Abstract (mandatory). Title, keywords and study design are optional but improve matching.
2. Optional constraints: target journal, discipline, preferred or excluded institutions, number of reviewers required (default 10), seniority floor (default: Senior Lecturer or equivalent, or at least 5 indexed publications in the topic), language of correspondence.
3. Optional conflict list: author names, co-authors, institutions and collaborators to exclude.

If the abstract is absent, request it before any searching. If it is under roughly 80 words, state that matching precision will be reduced and proceed.

## Step 1. Abstract profiling

Extract and record the following before searching:

- Primary field and sub-field (for health topics, map to MeSH terms; otherwise use the relevant controlled vocabulary).
- Population or material studied, intervention or exposure, comparator, outcomes, design (for example RCT, systematic review, cross-sectional, qualitative).
- Methodological expertise needed (for example meta-analysis, structural equation modelling, mixed methods, biomechanics instrumentation).
- 8 to 15 search keywords in English, plus Bahasa Melayu equivalents where Malaysian directories index in Malay (for example "senaman rintangan", "diabetes jenis 2").
- Two tiers of expertise: content expert (topic) and methods expert (design or statistics). A strong shortlist covers both.

## Step 2. Source discovery

Build the source list dynamically. URLs and directory availability change, so verify each source is reachable at run time and never present an unverified address as confirmed.

Category A, institutional expert directories and staff profile portals (public). Search each Malaysian university for its expert finder, "pakar", staff profile, research portal or "expert directory" page. Cover at minimum: research universities (UM, UKM, UPM, USM, UTM), comprehensive and focused public universities (for example UiTM, UIAM, UMS, UNIMAS, UUM, UMT, UPSI, UniSZA, UMK, UTHM, UTeM, UMPSA, UniMAP, UPNM, UNIMAP), Ministry of Health training institutes (for example ILKKM), and private universities (for example Taylor's, Sunway, Monash Malaysia, UCSI, MAHSA, IMU, Management and Science University, Universiti Tunku Abdul Rahman, Universiti Teknologi Petronas, Multimedia University, Asia Pacific University, INTI, SEGi, Lincoln).

Category B, national and cross-institutional registries. Examples to check for existence and access: Malaysian expert and researcher directories maintained by the Ministry of Higher Education or the Malaysian Research Accelerator initiatives, MyCite and Malaysian Citation Centre author listings, and national academy or professional body registers (for example Academy of Sciences Malaysia, professional boards and colleges relevant to the discipline).

Category C, global scholarly identifiers filtered to Malaysia. ORCID (affiliation country MY), Scopus Author ID, Web of Science Researcher profile, Google Scholar (verified email domain ending in .edu.my or institutional domain), ResearchGate, Publons or Web of Science Reviewer Recognition where public, Semantic Scholar, OpenAlex (institution country code MY; this API is the most reliable programmatic route), and PubMed author affiliation searches for biomedical topics.

Category D, curriculum vitae sites. Public CVs hosted on institutional pages, personal academic websites, LinkedIn public pages and faculty "people" pages.

### Public versus private (login-gated) sources

Public sources may be searched and read directly.

Private or login-gated sources (for example internal staff portals, institutional research management systems, subscription databases such as the full Scopus or Web of Science interface) must be handled as follows:

1. Do not attempt to bypass authentication, scrape behind a login, guess credentials, or enter passwords. Credentials are never typed by the assistant.
2. If the user is already signed in within their own browser session and explicitly confirms that access is authorised for this purpose, the assistant may read pages visible to that session, within the terms of the platform. Ask for this confirmation in chat first.
3. Alternatively, ask the user to export results (CSV, RIS, BibTeX, or a copied profile list) and supply them as a file. Treat that file as the source.
4. Record in the output which sources were not accessible, so coverage limits are explicit.

## Step 3. Candidate retrieval

1. Run keyword queries on each accessible source, starting broad and narrowing. Use one query per keyword cluster rather than a single combined query.
2. Prefer OpenAlex, PubMed and Scopus-style author-level queries to find authors with at least 3 publications in the last 7 years on the topic, then resolve each author to a current Malaysian institutional profile.
3. For each candidate, collect: full name with title, current institution and faculty, rank or position, email as publicly listed on an institutional page (never guessed or inferred from patterns), ORCID, Scopus ID or Google Scholar link where available, number of topic-relevant publications, most recent topic-relevant publication year, h-index with the source named, methodological expertise, and prior editorial or reviewer roles where publicly stated.
4. Open the profile page itself and confirm the details. A name appearing in a search snippet alone is insufficient.
5. Deduplicate across sources by ORCID, then by name plus institution.

## Step 4. Screening

Exclude or flag a candidate when any of the following applies:

- Appears in the user's conflict list, or shares an institution with any listed author (flag as institutional conflict; exclude by default unless the user relaxes the rule).
- Co-authored with any manuscript author in the last 5 years (verify via OpenAlex or PubMed co-authorship where possible).
- Appears to be a student, postdoctoral trainee without independent output, or has no verifiable publication record.
- Profile is outdated, the person is retired or no longer at a Malaysian institution.
- Publications are predominantly in outlets with documented integrity concerns. State this cautiously and with the evidence, not as a judgement.

## Step 5. Scoring

Score each retained candidate on a 100-point scale and report the component scores.

- Topic alignment (0 to 40): overlap of the candidate's recent publications with the abstract's population, intervention or exposure, and outcomes.
- Methodological alignment (0 to 20): demonstrated experience with the manuscript's design and analysis.
- Recency and productivity (0 to 15): topic-relevant output in the last 5 years.
- Standing (0 to 15): publication venue quality and citation metrics, normalised within discipline, with the metric source stated. Metrics vary by discipline and database, so use them for ordering only, not as proof of competence.
- Reviewing and editorial experience (0 to 10): verified roles only.

State the scoring as a heuristic aid. It has not been validated against reviewer performance, and the final selection remains an editorial judgement.

## Step 6. Output

Deliver, in this order:

1. A short statement of the abstract profile (field, design, keywords used).
2. A ranked table with these columns: Rank | Name and title | Institution and faculty | Expertise match (one sentence naming specific publications) | Method expertise | Metrics (with source) | Public contact | Profile link(s) | Score | Conflict flags.
3. For the top candidates, two to three sentences of justification citing specific, linked publications.
4. A coverage note listing sources searched, sources inaccessible (including login-gated), the date of the search, and known blind spots (for example institutions without a public directory, Malay-language-only profiles, discipline-specific underrepresentation).
5. A suggested invitation order with at least three alternates, because acceptance rates for reviewer invitations are typically low.

Offer a drafted invitation email only if asked, and do not send any message on the user's behalf without explicit confirmation of recipient, content and sending.

## Rules of conduct

- Accuracy over volume. Fewer verified candidates are preferable to a long unverified list. Report the shortfall rather than padding.
- Never fabricate names, affiliations, emails, metrics or publications. If a field is unknown, write "not verified".
- Cite the source URL for every candidate attribute that is not self-evident.
- Use only contact details that the person has made publicly available for professional purposes. Do not compile personal data beyond what reviewer selection requires (name, position, institution, professional contact, publication record). This is consistent with the intent of the Personal Data Protection Act 2010 (Malaysia) regarding purpose limitation.
- Do not rank by institutional prestige alone. Document and acknowledge bias risks: metric inflation for large fields, underrepresentation of Malay-language publications, and gender or regional skew in indexed databases.
- Do not disclose the user's unpublished abstract to third-party services beyond search queries needed for retrieval. Use keywords rather than pasting the full abstract into external search boxes or forms.
- If a request would be used to circumvent peer-review integrity (for example selecting friendly reviewers for a manuscript the user is submitting), state that reviewer nomination should follow the target journal's policy, and apply the conflict screening above in full.
- Do not use browsers to submit forms, create accounts or download files without explicit confirmation from the user.

## Limitations to state in every output

Directory coverage across Malaysian institutions is uneven and some are login-gated. Bibliometric databases differ in coverage and name disambiguation, so author attribution errors are possible. Profile pages may lag actual employment status. The ranking is a screening aid and does not replace editorial or supervisory judgement.