Intake Project Agent Skill

Intake-Project Skill

Purpose

This skill is the front door for a new or messy project. It converts a folder, document bundle, or mixed set of notes into a structured project state that other skills can use safely.

Use this skill when:

• a new paper or proposal folder has been created
• an older folder exists but is poorly organized
• the user asks "what is this project and what should I do next?"
• another skill needs a reliable project summary before proceeding

Communication Rules

• Communicate with the user in their preferred language.
• Keep project labels and file names in the language already used by the workspace.
• Use English for manuscript section names, study design names, and medical/statistical terminology.

Inputs

Accept any of the following:

• a project folder
• a manuscript draft
• an abstract or proposal
• tables/figures plus notes
• a mixed folder with PDFs, drafts, and analyses

If information is incomplete, infer cautiously from file names and contents, then label uncertain items clearly.

Core Tasks

1. Project classification

Determine:

• project type: original | review | meta-analysis | case report | technical note | grant | peer review | challenge | career-doc
• primary domain: radiology | medical AI | multimodal LLM | intervention | survival/prognostic | diagnostic accuracy | workflow
• target output: paper | abstract | grant | review | rebuttal | CV
• likely target journal or venue, if recoverable

2. State reconstruction

Identify:

• what already exists
• what is missing
• current phase
• blocking dependencies

3. Project memory scaffold

If missing, propose or create lightweight anchor files:

• PROJECT.md
• STATUS.md
• CLAIMS.md
• DATA_DICTIONARY.md
• ANALYSIS_PLAN.md
• REVIEW_LOG.md

Create only files that are justified by the project type.

4. Action plan

Produce the next 3-5 actions in dependency order.

Canonical Manuscript Folder Structure

For any manuscript project (cohort, MA, RCT, case series), enforce this structure when scaffolding or reorganizing. Map every new artifact into one of these slots — do not invent ad-hoc folders.

{project_root}/
├── HANDOFF.md                         # session handoff entry point
├── README.md                          # project overview
├── data/                              # raw data (NEVER edit; read-only)
├── analysis/                          # reproducible scripts (00_* → 04_*)
├── output/                            # analysis outputs: CSVs, PNGs, intermediates
├── irb/                               # IRB/ethics docs
├── proposal/                          # original protocol / approved proposal
├── reviews/                           # external correspondence
├── manuscript/                        # SOURCE manuscript + drafting
│   ├── manuscript_v{N}.{md,docx,pdf}  # current canonical working version (top level)
│   ├── build_unified_docx.py          # or pandoc wrapper
│   ├── archive/                       # ALL prior versions v1 .. v{N-1}
│   ├── reviews/                       # QC: self_review, peer_review, STROBE/PRISMA, critic
│   ├── figures/                       # figure scripts + rendered PNG/PDF
│   └── tables/                        # table scripts + rendered docx
└── submission/                        # per-journal packages
    └── {journal-slug}/                # e.g., chest/, kjr/
        ├── CHECKLIST.md
        ├── cover_letter.{md,docx,pdf}
        ├── title_page.docx            # separated for double-anonymized
        ├── manuscript_anonymized.{docx,pdf}
        ├── supplement.{docx,pdf}
        ├── strobe_checklist.md        # or PRISMA / CONSORT
        ├── circulation_email.md
        └── figures/                   # submission-ready DPI copies

Rules

• manuscript/ = source; submission/{journal}/ = derived artifacts. Regenerate submission files from manuscript/manuscript_v{N}.md; never edit anonymized/title-page directly.
• One canonical working version at manuscript/manuscript_v{N}.{md,docx,pdf}. Older versions move to manuscript/archive/ immediately on version bump.
• No loose files at project root. Only HANDOFF.md, README.md, folder entries.
• QC artifacts (self_review, peer_review, STROBE, critic reports) live in manuscript/reviews/, not at manuscript top level.
• On rejection/retarget: cp -r submission/{old} submission/{new}, then rewrite cover letter and reformat.
• Double-anonymized journals (Chest, AJRCCM): title page and anonymized manuscript MUST be separate files under submission/{journal}/.

When to apply

• At project intake: scaffold empty structure.
• At first submission prep: create submission/{journal}/ and populate.
• Mid-project cleanup: when manuscript/ has >3 versioned files or QC docs at top level, reorganize.
• Before session handoff: reorganize if structure is drifting.

Precedent: an STROBE cohort with a mortality endpoint reorganized v1–v6 plus QC docs from the manuscript/ top level so that a reject-retarget path to a different journal requires only cp -r submission/chest submission/<new_journal>.

Workflow

Phase 1: Discover context

• Read top-level folder names and key files.
• Detect manuscript-like files, tables, figures, protocols, and analysis outputs.
• Extract:
• project title or working title
• study question
• dataset or cohort hints
• collaborators or institutions
• venue/journal hints

Phase 2: Classify project stage

Assign one current stage:

• idea
• data assembly
• analysis planning
• analysis in progress
• drafting
• revision
• submission prep
• archived/unclear

Gate: Present the classification (project type, stage, target output) to the user.

Confirm before creating any files — misclassification leads to wrong scaffold and

wrong skill routing.

Phase 3: Surface missing inputs

Check for common gaps:

• no explicit study question
• no target journal
• no analysis plan
• no variable dictionary
• no claims-to-results map
• no review log for revised manuscripts

Phase 4: Produce normalized summary

Output this structure:

## Project Intake Summary
Project: ...
Type: ...
Current stage: ...
Likely target: ...

### What exists
- ...

### What is missing
- ...

### Risks / ambiguities
- ...

### Recommended next actions
1. ...
2. ...
3. ...

Optional File Templates

PROJECT.md

# PROJECT

- Title:
- Type:
- Primary question:
- Target journal/venue:
- Lead folder:
- Collaborators:
- Last updated:

STATUS.md

# STATUS

- Current stage:
- Current blocker:
- Next actions:
  1.
  2.
  3.
- Last updated:

Guardrails

• Do not invent data values, outcomes, or collaborator roles.
• Do not assume a target journal unless evidence exists in the files.
• Do not create a large folder scaffold when the user only wants a quick assessment.
• If a project appears to mix multiple studies, say so explicitly rather than collapsing them into one.

Handoff Rules

After intake:

• route to search-lit if the literature basis is weak
• route to design-study if the research question exists but design logic is unclear
• route to manage-project if the folder should be scaffolded
• route to write-paper only after the project phase is clearly drafting

What This Skill Does NOT Do

• It does not write full manuscript sections
• It does not perform statistical analysis
• It does not verify citations deeply
• It does not replace study design review

Anti-Hallucination

• Never fabricate file paths, URLs, DOIs, or package names. Verify existence before recommending.
• Never invent journal metadata, impact factors, or submission policies without verification at the journal's website.
• If a tool, package, or resource does not exist or you are unsure, say so explicitly rather than guessing.