How do I install the Write Report skill?

Run npx skillmds@latest add majiayu000/write-report in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Write Report skill do?

Generates a human-readable markdown review report from an xreview session. Invoked automatically by the xreview skill at Step 5 after a review is complete, or manually via /write-report . Combines structured session data with conversation context (verification results, user decisions, fix details) to produce a comprehensive, readable report. It is listed under Docs & Writing on SkillMD.

Is Write Report safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Write Report?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Write Report free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Write Report?

majiayu000 (@majiayu000) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Write Report by majiayu000 — Agent Skill

Name: Write Report
Rating: 4.2 (5 reviews)
Author: majiayu000

Write Report

Generates a human-readable markdown review report from an xreview session. Invoked automatically by the xreview skill at Step 5 after a review is complete, or manually via /write-report <session-id>. Combines structured session data with conversation context (verification results, user decisions, fix details) to produce a comprehensive, readable report.

majiayu000 Updated Sep 9, 2026 567 repo stars

File contents

write-report — Generate Review Report

Step 1: Get session data

Run: xreview report --session <session-id>

Parse the XML output to extract:

Session ID and round count
Finding list with final statuses (open / fixed / dismissed)
Summary counts (total, open, fixed, dismissed)
Report file path (the JSON data xreview saved)

If the command fails, show the error to the user and stop.

Step 2: Generate markdown report

Combine the structured session data with what you recall from the conversation:

Verification results — which findings you confirmed or suspected, and why
User decisions — what the user chose to fix, skip, or adjust
Fix details — what code changes you made for each finding
Dismissal reasons — why certain findings were not fixed

Write the report using this layout:

# Code Review Report

**Date:** YYYY-MM-DD
**Session:** <session-id>
**Reviewed:** <list of files or "uncommitted changes">
**Rounds:** <number of review rounds>

## Context

<The --context provided for this review.
Include what was reviewed, why, and what the reviewer should focus on.>

## Summary

| Status | Count |
|--------|-------|
| Total findings | N |
| Fixed | N |
| Dismissed | N |
| Open | N |

## Findings

### F-XXX: <title> — <severity>/<category> ✅ Fixed

**Location:** `file.go:42`
**Issue:** <description of the problem>
**Fix applied:** <what was changed and why this approach was chosen>

### F-XXX: <title> — <severity>/<category> ⏭️ Dismissed

**Location:** `file.go:10`
**Issue:** <description>
**Reason:** <why — user decision, false positive confirmed by Codex discussion, etc.>

### F-XXX: <title> — <severity>/<category> ⚠️ Open

**Location:** `file.go:99`
**Issue:** <description>
**Note:** <why this remains open — deferred to next PR, needs more context, etc.>

## Open Items

<If any findings remain open, summarize what's left and recommended next steps.
If everything is resolved, write "All findings have been addressed.">

Layout rules:

Group findings by status: Fixed first, then Dismissed, then Open
For Fixed findings: describe what was actually changed, not just the original suggestion
For Dismissed findings: always include the reason
Use emoji status markers consistently: ✅ Fixed, ⏭️ Dismissed, ⚠️ Open
Keep descriptions concise — the report is a record, not a tutorial

Step 3: Save and present

Save the full report to .xreview/reports/<session-id>.md using the Write tool.
Print a summary in the conversation:

Review complete. Report saved to .xreview/reports/<session-id>.md

Summary: N findings — X fixed, Y dismissed, Z open.

Clean up session — run xreview clean --session <session-id>. The session data is no longer needed after the report is saved.

---name: write-reportdescription: Generates a human-readable markdown review report from an xreview session. Invoked automatically by the xreview skill at Step 5 after a review is complete, or manually via /write-report <session-id>. Combines structured session data with conversation context (verification results, user decisions, fix details) to produce a comprehensive, readable report.--- # write-report — Generate Review Report ## Step 1: Get session data Run: `xreview report --session <session-id>` Parse the XML output to extract: - Session ID and round count - Finding list with final statuses (open / fixed / dismissed) - Summary counts (total, open, fixed, dismissed) - Report file path (the JSON data xreview saved) If the command fails, show the error to the user and stop. ## Step 2: Generate markdown report Combine the structured session data with what you recall from the conversation: - **Verification results** — which findings you confirmed or suspected, and why - **User decisions** — what the user chose to fix, skip, or adjust - **Fix details** — what code changes you made for each finding - **Dismissal reasons** — why certain findings were not fixed Write the report using this layout: ```markdown # Code Review Report **Date:** YYYY-MM-DD **Session:** <session-id> **Reviewed:** <list of files or "uncommitted changes"> **Rounds:** <number of review rounds> ## Context <The --context provided for this review. Include what was reviewed, why, and what the reviewer should focus on.> ## Summary | Status | Count | |--------|-------| | Total findings | N | | Fixed | N | | Dismissed | N | | Open | N | ## Findings ### F-XXX: <title> — <severity>/<category> ✅ Fixed **Location:** `file.go:42` **Issue:** <description of the problem> **Fix applied:** <what was changed and why this approach was chosen> ### F-XXX: <title> — <severity>/<category> ⏭️ Dismissed **Location:** `file.go:10` **Issue:** <description> **Reason:** <why — user decision, false positive confirmed by Codex discussion, etc.> ### F-XXX: <title> — <severity>/<category> ⚠️ Open **Location:** `file.go:99` **Issue:** <description> **Note:** <why this remains open — deferred to next PR, needs more context, etc.> ## Open Items <If any findings remain open, summarize what's left and recommended next steps. If everything is resolved, write "All findings have been addressed."> ``` **Layout rules:** - Group findings by status: Fixed first, then Dismissed, then Open - For Fixed findings: describe what was actually changed, not just the original suggestion - For Dismissed findings: always include the reason - Use emoji status markers consistently: ✅ Fixed, ⏭️ Dismissed, ⚠️ Open - Keep descriptions concise — the report is a record, not a tutorial ## Step 3: Save and present 1. **Save the full report** to `.xreview/reports/<session-id>.md` using the Write tool. 2. **Print a summary in the conversation:** ``` Review complete. Report saved to .xreview/reports/<session-id>.md Summary: N findings — X fixed, Y dismissed, Z open. ``` 3. **Clean up session** — run `xreview clean --session <session-id>`. The session data is no longer needed after the report is saved.

Frequently asked questions