text-reference-mapper

name: text-reference-mapper description: Map verifiable academic references onto user-written text for management research. Takes a passage of text, extracts key claims, finds 1–2 peer-reviewed citations per claim from top management journals, and outputs a structured evidence map, formatted bibliography, and gap analysis. Use this skill whenever the user says "reference this text", "find citations for my writing", "add references to this paragraph", "cite this section", "evidence map", "support these claims with references", "back up my arguments", or provides a block of academic writing and asks for supporting citations. Also triggers when users request a specific citation style (APA 7, Harvard, Chicago) for their own text. This skill is specifically designed for management scholarship—strategy, entrepreneurship, innovation, organizational behavior, international business, and adjacent fields.

Text Reference Mapper

Take user-written academic text, extract its key claims, and map each claim to verified peer-reviewed references from top management journals. The output is always a structured evidence map, a formatted bibliography, and a gap analysis.

This skill differs from general paper-finding: it starts with the user's own text, not a topic query. The goal is to provide a verifiable scholarly backbone for an existing argument.

Core Principle

Accuracy over completeness. A single verified citation is worth more than ten plausible-sounding fabrications. Fabricated citations can end academic careers. Every reference must be independently verifiable.

Anti-Hallucination Rules

These rules are non-negotiable and override any temptation to be "helpful" by guessing:

1. Never fabricate citations. If you are not 100% sure a paper exists based on search results, do not include it.
2. Every citation must be independently verifiable using at least one of: publisher/journal article page, Crossref DOI record, Web of Science record, or Scopus record.
3. Search before citing. Never rely on training data for specific paper details (titles, authors, years, volumes, pages).
4. Use exact details from search results. Never reconstruct titles, author names, or years from memory.
5. If you cannot verify, say so. Gaps are valuable information—flag them honestly.

Inputs

The user provides:

If the user doesn't specify a citation style, ask once. If they don't specify optional parameters, proceed with defaults.

Procedure

Step 1: Extract Claims

Read the user's text carefully. Extract up to 12 key claims—statements that assert something empirically testable, theoretically grounded, or causally directional. Keep each claim to one sentence.

What counts as a claim:

• Causal assertions: "X leads to Y", "X improves Y"
• Theoretical propositions: "According to [theory], X should..."
• Empirical statements: "Research shows that...", "Studies have found..."
• Definitional anchors: "X is defined as..."
• Boundary conditions: "This effect is stronger when..."

What is NOT a claim (skip these):

• Pure description of the paper's own methodology
• Transition sentences
• Statements of intent ("This paper examines...")

Present the extracted claims to the user as a numbered list before proceeding to search. This lets the user correct misinterpretations early.

Step 2: Search and Verify

For each claim, search for 1–2 supporting peer-reviewed articles. Follow this search strategy:

Search sequence:

1. Concept + top journal search: "[key concept]" "Academy of Management" OR "Strategic Management Journal" OR "Organization Science"
2. Google Scholar search: "[key concept]" [related terms] site:scholar.google.com
3. Author-targeted search (if a theoretical tradition is obvious): [known author] "[concept]" [journal]
4. DOI verification: Once a paper is found, search for its DOI on doi.org or Crossref to confirm metadata

Preferred outlets (use when relevant):

• Academy of Management Journal (AMJ)
• Academy of Management Review (AMR)
• Administrative Science Quarterly (ASQ)
• Strategic Management Journal (SMJ)
• Organization Science
• Journal of Management
• Journal of Management Studies
• Management Science
• Organization Studies

Field-adjacent top journals are acceptable when the core management outlets don't cover the claim (e.g., Research Policy for innovation, Journal of Business Venturing for entrepreneurship, Journal of International Business Studies for IB).

Verification protocol—for every paper you cite, confirm:

Red flags—do not cite if:

• You're reconstructing details from memory rather than search results
• Author name "sounds right" but doesn't appear in current search results
• You cannot find the paper when searching its exact title
• Journal name or year is estimated rather than confirmed

Step 3: Build the Evidence Map

Produce a table mapping each claim to its supporting references. Include a 1–2 sentence rationale explaining why the citation supports the claim.

Evidence map format:

Step 4: Produce the Reference List

Format all cited works in the user's requested citation style. For each reference, include:

[Formatted citation in requested style]
DOI: https://doi.org/xx.xxxx/xxxxx
Verification: [Crossref DOI record / Publisher page / WoS / Scopus]

If no DOI is available:

[Formatted citation]
No DOI listed. Publisher page: [URL]
Verification: [Publisher page]

Citation style reference:

Read references/citation-styles.md for the exact formatting rules for APA 7, Harvard, and Chicago author-date styles.

Step 5: Gap Analysis

After completing the evidence map, report:

1. Unsupported claims: Claims for which no strong journal support was found
2. Weak matches: Claims where the citation is tangential rather than direct
3. Alternative source suggestions: For claims needing industry reports, government statistics, or practitioner sources—specify what kind of source is needed but do not cite non-peer-reviewed sources unless the user explicitly allows it
4. Recommended search keywords: For each gap, suggest 3–5 search terms the user could try in Google Scholar or Web of Science

Output Structure

Always deliver three sections in this order:

Section 1: Evidence Map

The claim-to-citation table from Step 3.

Section 2: Reference List

The formatted bibliography from Step 4, sorted alphabetically by first author surname.

Section 3: Gaps & Suggestions

The gap analysis from Step 5.

Handling Edge Cases

User provides very short text (1–2 sentences): Extract 1–3 claims and proceed normally. Flag if the text is too vague for targeted searching.

User's claims are highly specific or niche: Say so directly: "This claim is very specific and may require [type of source]. I was unable to find direct support in top management journals." Suggest adjacent search strategies.

User requests non-management references: This skill is optimized for management scholarship. For other fields, acknowledge the limitation and suggest the user specify relevant journals for that field.

User provides text with existing citations: Ask whether they want to verify existing citations, find additional citations, or replace weak ones. Then proceed accordingly.

Conflicting evidence found: Report it. If a claim is contested in the literature, note both supporting and contradicting references. The user benefits from knowing this.

Dependencies

• web_search: Required for finding papers (always use before citing)
• web_fetch: Use to verify paper details from publisher/journal pages and Crossref

Quality Checklist

Before delivering the final output, confirm:

• Every citation was found via web search (not from memory)
• Every title is copied verbatim from a search result or publisher page
• Every DOI resolves correctly (or is marked "No DOI listed")
• Every reference includes a Verification line
• No reference has estimated or reconstructed metadata
• Gap analysis honestly flags unsupported claims
• Citation style is consistent throughout