dh-review

dh-review — Feedback & Review

What This Skill Does

dh-review provides two types of assistance:

1. Review simulation: Evaluates your DH paper from three perspectives — technical (methodology, data, reproducibility), humanistic (argument, scholarship, interpretation), and interdisciplinary (the DH bridge). Also includes devil's advocate challenge.

2. Feedback parsing: Transforms unstructured supervisor, advisor, or peer reviewer comments into a prioritized, actionable revision roadmap.

Who It's For

• Graduate students preparing to submit a paper and wanting pre-submission review
• Researchers who have received supervisor or reviewer feedback and need help prioritizing the revision
• Anyone wanting to stress-test their DH argument before exposing it to real reviewers

---

Modes and Trigger Keywords

Mode	Trigger	What Happens
full-review	/dh-review or /dh-review full-review	Complete multi-perspective review (all 4 review agents)
supervisor-feedback	/dh-review supervisor-feedback	Parse supervisor/reviewer comments → revision roadmap
technical-focus	/dh-review technical-focus	Computational methodology and reproducibility review only
humanities-focus	/dh-review humanities-focus	Humanistic argument and scholarly engagement review only
interdisciplinary	/dh-review interdisciplinary	DH bridge review only — does the paper justify being DH?
quick-assess	/dh-review quick-assess	Rapid high-level assessment; ~10 minutes

---

Mode Routing

1. Always start with intake: Read agents/intake_agent.md to detect input type and configure review balance.

2. Route by mode:

   • full-review: Execute all four review agents in sequence, then compile the Reviewer Decision
   • supervisor-feedback: Execute supervisor_feedback_agent.md only
   • technical-focus: Execute technical_reviewer_agent.md only
   • humanities-focus: Execute humanities_reviewer_agent.md only
   • interdisciplinary: Execute interdisciplinary_reviewer_agent.md only
   • quick-assess: Execute all four agents; limit each to 150 words; compile a condensed report

---

Review Calibration

Before the review, read references/reviewer_calibration_guide.md and set the technical/humanistic weight ratio based on the target venue. The balance shifts from 35/65 (DHQ) to 65/35 (Cultural Analytics).

---

The Reviewer Decision (Full Review Output)

After full-review, produce a single consolidated output:

## Reviewer Decision

**Paper**: [title]
**Target Venue**: [venue]
**Review Date**: [date]

**Overall Recommendation**: [Accept (minor revisions) | Minor Revisions | Major Revisions | Reject/Resubmit]

**Scores**:
| Criterion | Score (1–5) |
|---|---|
| Research Question Quality | [n] |
| Corpus Integrity | [n] |
| Methodological Soundness | [n] |
| Interpretive Depth | [n] |
| Scholarly Engagement | [n] |
| DH Bridge | [n] |
| **Overall** | **[n]** |

**Summary**: [2–3 paragraph integrated review — synthesis of all three reviewer perspectives]

**Revision Roadmap**: [Tier 1 / Tier 2 / Tier 3 items, consolidated from all reviewers]

---

Handoff

After full-review: the Reviewer Decision + Revision Roadmap is detected by dh-write's intake agent when passed into a revision mode session.

After supervisor-feedback: the Revision Roadmap is detected by dh-write revision mode.

In dh-pipeline, Stage 5.5 uses this skill's supervisor-feedback mode to process advisor feedback between review rounds.