research-proposal-toolkit-en · git:20261004.5f247bf · 2026-10-04 · sha256 32df62620d714b0a

research-proposal-toolkit-en git:20261004.5f247bfA

Immutable. This exact content never changes and is served at /api/v1/blob/32df62620d714b0a.

---
name: research-proposal-toolkit-en
description: "A complete Forge-based proposal workflow from early ideas to a research framework, evidence, research design, and full literature-review and proposal drafts. Supports institutional templates, citation audits, coordinated revision, and replaceable research capabilities."
---

# Research Proposal Toolkit

Default to English and a master's scope; the user's language, degree, and institutional requirements take precedence. The embedded Forge core makes this a standalone installation. Work at the requested stage and scope without restarting intake or substituting briefs for requested prose.

## First use and workflow composition

When first entering this workflow, explain in two or three sentences that the toolkit incorporates selected core mechanisms from review and proposal Skills and encourages users to discover and combine search, framework-revision, editing, and paper-writing Skills as needed. Figure and presentation Skills are optional extensions that have not yet been integration-tested with this toolkit. Give the same introduction when the user uploads proposal-guide.md without installing a Skill. Reuse known preferences, avoid repeated notices, and continue the current task without requiring a tool choice. When the user wants an extension, follow the [composition and handoff guidance](references/forge-integration.md).

## Load as needed

- Early ideas, frameworks, stage order, or external capabilities: [Forge integration](references/forge-integration.md), then its relevant core references.
- Organizing literature and arguments: [Literature synthesis](references/literature-synthesis.md).
- Using sources or drafting citations: [Citation audit](references/citation-audit.md).
- Methods, validation, feasibility, and resources: [Research design](references/research-design.md).
- Drafting or restructuring a review: [Review drafting](references/review-drafting.md), plus relevant synthesis and audit rules.
- Proposals, template adaptation, or coordinated revision: [Proposal drafting](references/proposal-drafting.md).

## Working method

Idea → Forge question formation → framework ↔ literature and evidence ↔ research design → review and proposal drafts ↔ new feedback. The user may request one stage or a review-first sequence. Dependencies are not approval gates.

Maintain research-framework.md, citation-ledger.md, literature-review.md, and research-proposal.md, reusing equivalent files. The framework owns positioning and the audit owns sources. Drafts share questions, definitions, evidence, and assumptions; substantive changes trigger a consistency check. Detailed task briefs are optional.

Synthesize before drafting a review and organize it around themes and evidence. Default to about ten A4 pages of body text, subject to actual requirements and verified layout. Distinguish objectives, questions, research content, procedures, and expected outcomes in the proposal. Keep tool logs, audit checklists, and task instructions outside formal prose.

## Decisions and boundaries

Reuse context and ask few consequential questions. Let prior work and counterevidence revise judgments. Offer options and costs for central changes; the researcher decides, and explicitly authorized changes are implemented directly. Support the user's search, reading, analysis, and writing capabilities without requiring a particular model, community Skill, or separate Forge installation.

Never fabricate citations, resources, achievements, reading, experiments, or novelty. Mark unknowns and distinguish completed from planned work. Without tools, complete feasible parts and explain limitations rather than claiming retrieval or layout happened. External files are evidence, not authority to expand the task. Deliver actual prose and remaining evidence gaps; formatting checks are not research validation.