GitHub Review Iteration
prisma/orm
Runs a loop on a GitHub pull request: fetch review state, triage comments into actions, implement them and resolve threads, repeating until nothing actionable is left.
Delegate code review to a subagent running a specific model.
Install Cross Review
$ npx skills add EliasOulkadi/shokunin --skill cross-review -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install EliasOulkadi/shokunin cross-review --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/EliasOulkadi/shokunin.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.pack/skills/cross-review .claude/skills/cross-review && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "cross-review" agent skill from https://github.com/EliasOulkadi/shokunin/tree/master/.pack/skills/cross-review into .claude/skills/cross-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cross-review", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/EliasOulkadi/shokunin/tree/master/.pack/skills/cross-reviewType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add EliasOulkadi/shokunin --skill cross-review -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install EliasOulkadi/shokunin cross-review --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/EliasOulkadi/shokunin.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.pack/skills/cross-review .agents/skills/cross-review && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "cross-review" agent skill from https://github.com/EliasOulkadi/shokunin/tree/master/.pack/skills/cross-review into .agents/skills/cross-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cross-review", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add EliasOulkadi/shokunin --skill cross-review -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install EliasOulkadi/shokunin cross-review --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/EliasOulkadi/shokunin.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.pack/skills/cross-review .cursor/skills/cross-review && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "cross-review" agent skill from https://github.com/EliasOulkadi/shokunin/tree/master/.pack/skills/cross-review into .cursor/skills/cross-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cross-review", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/EliasOulkadi/shokunin.git --path .pack/skills/cross-review--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add EliasOulkadi/shokunin --skill cross-review -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install EliasOulkadi/shokunin cross-review --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/EliasOulkadi/shokunin.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.pack/skills/cross-review .gemini/skills/cross-review && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "cross-review" agent skill from https://github.com/EliasOulkadi/shokunin/tree/master/.pack/skills/cross-review into .gemini/skills/cross-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cross-review", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install EliasOulkadi/shokunin cross-reviewInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add EliasOulkadi/shokunin --skill cross-review -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/EliasOulkadi/shokunin.git skills-src && mkdir -p .github/skills && cp -r skills-src/.pack/skills/cross-review .github/skills/cross-review && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "cross-review" agent skill from https://github.com/EliasOulkadi/shokunin/tree/master/.pack/skills/cross-review into .github/skills/cross-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cross-review", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add EliasOulkadi/shokunin --skill cross-review -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install EliasOulkadi/shokunin cross-review --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/EliasOulkadi/shokunin.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.pack/skills/cross-review .opencode/skills/cross-review && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "cross-review" agent skill from https://github.com/EliasOulkadi/shokunin/tree/master/.pack/skills/cross-review into .opencode/skills/cross-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cross-review", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
cross-review- GitHub stars
- 114
- Token cost
- ~2.2k tokens
- SKILL.md length
- 949 words
- Files
- 1
- Skills in repo
- 49
- Repo updated
- First seen
- Licence
- MIT
At a glance
Delegate code review to a subagent running a specific model.
- Works in 4 steps: Parse the user request → Gather context from your own changes → Spawn the review subagent → …
- Explicitly names a model to review changes (review with opus
- SKILL.md covers Workflow, Model Mapping, Error Handling and Anti-Patterns, plus 3 more sections
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Cross Review is an agent skill from EliasOulkadi/shokunin. Delegate code review to a subagent running a specific model. Use ONLY when user explicitly names a model to review changes ("review with opus", "use sonnet to review", "review with gemini"). The root agent reconstructs changes from conversation history and spawns a subagent with the code-review skill using the specified model. Do NOT use for general code review (use code-review skill instead), for reviewing PRs from git history, or when no model is specified by the user.
Its SKILL.md is about 2.2k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts. Compatibility notes: opencode
It sits in Development, covering Subagents, Git workflow and Code review. The repository describes itself as: 職人 Shokunin 62 AI agent skills for OpenCode, Claude Code, Cursor, Windsurf. ChromaDB memory, MCP servers, declarative self-updates. Multi-model, open source, zero cost. The licence is MIT.
When your agent uses it
- Explicitly names a model to review changes (review with opus
- Use sonnet to review
- Review with gemini)
- General code review (use code-review skill instead)
Example prompts
- “review with opus”
- “use sonnet to review”
- “review with gemini”
- “/cross-review”
Requirements
- Compatibility (from SKILL.md): opencode
Workflow steps
4 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit 4c68e5b. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Compatibility
opencode
From compatibility in the SKILL.md frontmatter.
Context cost
Cross Review loads about 2.2k tokens when it runs. Until then it costs about 122 tokens; SKILL.md has 949 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from EliasOulkadi/shokunin at commit 4c68e5b, republished under its MIT licence (© EliasOulkadi). 949 words, ~2,161 tokens.
Download SKILL.mdSave it as
.claude/skills/cross-review/SKILL.md (or your agent's skills folder).- name
- cross-review
- description
- Delegate code review to a subagent running a specific model. Use ONLY when user explicitly names a model to review changes ("review with opus", "use sonnet to review", "review with gemini"). The root agent reconstructs changes from conversation history and spawns a subagent with the code-review skill using the specified model. Do NOT use for general code review (use code-review skill instead), for reviewing PRs from git history, or when no model is specified by the user.
- compatibility
- opencode
- triggers
- review with, review using, review via, cross review, different model review
- negatives
- comprehensive review, zen review, simple review, code review
- license
- MIT
- metadata.workflow
- review
- metadata.audience
- developers
- metadata.version
- 1.0.0
Reconstruct what you changed during this conversation, then delegate the actual review to a single subagent running the code-review skill with a user-specified model.
IMPORTANT: Steps 1-2 run in the current agent (the master). Only Step 3 spawns a subagent.
Workflow
Step 1: Parse the user request
Extract:
- Model: The model ID from user's request. Validate against available models.
- Review instructions: Any text after "Review instructions:" — pass verbatim.
- Change scope: What should be reviewed. Default: all changes in this conversation.
Step 2: Gather context from your own changes
Reconstruct the diff from conversation history:
Compose a unified diff of all changes (Edit, Write, Bash, etc.). Group by file.
- If no changes: Inform user and stop.
Read final state of changed files for surrounding context.
Check related context:
- Test files related to changed files (skip
node_modules,.git,dist,build) - Config changes that affect behavior
- Related type definitions or interfaces
- Test files related to changed files (skip
Do NOT show gathered context to user. Use only for the subagent.
Step 3: Spawn the review subagent
spawn_subagent:
skill: "code-review"
model: <model from user request>
prompt: |
Review the changes below using "code-review" skill.
CONSTRAINTS:
- Read-only review. Do NOT edit files.
- All context is provided below. Read files only if clearly incomplete.
- Review independently and objectively.
## Review Instructions
{user's review instructions, verbatim}
## Changes
{reconstructed diff, grouped by file}
## Additional Context
{links to related files, tests, type definitions, requirements}Step 4: Relay the result — READ-ONLY, NO ACTIONS
CRITICAL: Output the subagent's review AS-IS. Do NOT:
- Summarize, rephrase, reorder, or filter
- Fix, improve, or refactor based on findings
- Add your own commentary or caveats
One-line model attribution is acceptable. You may offer to implement recommendations, but let the user decide.
Model Mapping
When user says "review with [model name]", map to a valid model ID:
| User says | Model ID |
|---|---|
| opus, claude opus | claude-3-opus |
| sonnet, claude sonnet | claude-3.5-sonnet |
| gpt-4, gpt4 | gpt-4 |
| gpt-4o | gpt-4o |
| gemini | gemini-2.0-flash |
| deepseek | deepseek-chat |
If the model name is unrecognized, ask for the exact model ID.
Error Handling
- Subagent fails or times out: Inform user. Suggest retry or different model.
- No changes in conversation: Inform user and stop.
- Incomplete review: Relay what was returned. Note it may be incomplete.
- Model not available: Offer alternative from the mapping table.
- User requests unknown model: Map fuzzy names to model IDs using the mapping table. If no match, ask user to provide the exact model ID string.
- Subagent returns malformed output: The review subagent may return text instead of structured findings. Relay what was returned and note the format deviation.
- Diff context exceeds subagent limits: For very large changesets, split the diff by file or module and run multiple sequential subagent reviews. Inform user you are splitting the review.
- Network or infrastructure failure during spawn: Retry once with the same configuration. If it fails again, offer to switch to a different model or perform the review inline.
- Subagent produces factually incorrect findings: Relay findings as-is but append a note that some claims could not be verified against the codebase. Do not filter or censor.
Anti-Patterns
| Pattern | Problem | Fix |
|---|---|---|
| Using cross-review for general PR review | User did not name a model. cross-review is for explicit model-named delegation only. | Use the standard code-review skill instead. |
| Summarizing or filtering subagent output | The user requested an independent review. Any distortion defeats the purpose. | Relay output verbatim. Add only a 1-line model attribution. |
| Acting on findings without user approval | The master agent is read-only in this workflow. Fixing issues automatically breaks the delegation contract. | Offer to implement recommendations but wait for user to decide. |
| Skipping context gathering (Step 2) | Sending only the diff without file context produces shallow reviews that miss type errors and behavioral changes. | Always read final file state and related test/config files before spawning the subagent. |
| Delegating to the same model the user is already talking to | If user says "review with sonnet" and master agent IS sonnet, this is circular. | Use a different model than the one running the master agent. |
| Requesting review of unchanged code | cross-review only works on changes made in the current conversation. | If the user wants PR review from git, use the code-review skill directly. |
Show full SKILL.md (323 more words)Show less
Model Comparison Methodology
When to use which model
| Model | Strengths | Best for |
|---|---|---|
| Claude Opus 4 | Deep reasoning, security analysis, architectural review | Complex refactors, auth code |
| GPT-5 Codex | Code generation patterns, syntax accuracy | Implementation review, template checking |
| Gemini 2.5 Pro | Multi-modal, context window size | Large PRs (200+ files), full-repo context |
| DeepSeek V4 | Fast, cost-effective, strong at bug detection | Routine PRs, quick checks |
Confidence Scoring
Assign confidence (0.0-1.0) to each finding:
confidence = (models_agreeing / total_models) * evidence_score
evidence_score:
1.0 = exact line + code excerpt proves the bug
0.7 = logic analysis suggests probable issue
0.4 = speculative (pattern matching without code walk)
0.1 = style preferenceFindings with confidence < 0.5 should be:
- Marked as "suggestion" not "finding"
- Never block a merge
- Phrased as questions: "Did you consider...?"
Result Merging Strategy
- Group by file + line range (+-5 lines)
- If all 3 models agree: high confidence, use as primary finding
- If 2 of 3 agree: medium confidence, note the dissenting model's view in details
- If 1 of 3: low confidence, verify against code manually before including
- Merge descriptions: take the most specific one, append key insights from others
- Remove duplicates: same finding described differently -> keep clearest description
Handling Conflicts
When models disagree:
- Don't default to majority -- investigate the code yourself
- The dissenting model may have caught something the others missed
- Present both views: "Claude suggests X, Gemini suggests Y"
- Let the human reviewer decide
Checklist
- Valid model name passed (not an alias or unsupported ID)
- Diff or context reconstructed accurately from conversation history
- Subagent has the code-review skill loaded
- Review result captured from subagent output (not assumed)
- Fallback model specified if primary model is unavailable
Sources
- MCP Protocol specification (modelcontextprotocol.io) — subprocess agent spawning and message passing
- OpenAI API model list documentation (platform.openai.com/docs/models) — supported model IDs and capabilities
- Anthropic Claude model documentation (docs.anthropic.com/en/docs/about-claude/models) — model IDs and feature comparison
- Google Gemini API model documentation (ai.google.dev/models/gemini) — Gemini model identifiers and rate limits
- DeepSeek API documentation (platform.deepseek.com/api-docs) — model IDs and context window limits
- Conventional Comments specification (conventionalcomments.org) — structured review comment formatting
- "Software Engineering at Google" by Titus Winters, Tom Manshreck, Hyrum Wright (O'Reilly, 2020) — code review best practices at scale
© EliasOulkadi, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in .pack/skills/cross-review of EliasOulkadi/shokunin.
Open the folder on GitHubat commit 4c68e5b
Compare with similar skills
Cross Review next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Cross Review this skillEliasOulkadi/shokunin | 114 | — | ~2.2k | Automated safety check: Pass | MIT | |
| GitHub Review Iterationprisma/orm | 48k | — | ~2.2k | Automated safety check: Pass | Apache-2.0 | |
| Cherry Studio PR ReviewCherryHQ/cherry-studio | 53k | — | ~3.9k | Automated safety check: Pass | AGPL-3.0 | |
| Jevgrepdzhng/jevgrep | 2.5k | — | ~741 | Automated safety check: Pass | MIT | |
| PR Reviewjaemk/self_update | 961 | — | ~1.5k | Automated safety check: Notes | MIT | |
| Cursor Composer Task DelegateChachamaru127/claude-code-harness | 3.2k | — | ~4.4k | Automated safety check: Notes | MIT |
Similar skills
- Official
Cherry Studio PR Review
CherryHQ/cherry-studio
Reviews Cherry Studio branches, pull requests, commits, files and docs against the project's own architecture, naming, API-boundary and UI rules, report-only by default.
Jevgrep
dzhng/jevgrep
A skill your agent uses for questions about how, why, or where behavior works in a repository, including questions that name a function or setting.
PR Review
jaemk/self_update
Targeted, read-only review of a PR or checked-out branch. An agent skill from jaemk/self_update.
Cursor Composer Task Delegate
Chachamaru127/claude-code-harness
Hands one implementation task to Cursor Composer in an isolated git worktree, then reviews its diff and cherry-picks the result into the main branch.
Local PR Review
windmill-labs/windmill
Runs the same code review locally that GitHub's auto-review actions run on a PR, delegating to a fresh-context subagent so the review isn't biased by the main session's own reasoning.
More from EliasOulkadi/shokunin
All 49 skills in this repoCI CD
EliasOulkadi/shokunin
Design CI/CD pipelines for GitHub Actions, GitLab CI, and CircleCI with matrix builds, test sharding, caching, Docker layer caching, OIDC auth, deployment strategies (rolling, blue-green, canary)…
Component Forge
EliasOulkadi/shokunin
Build production-grade components for React, Vue 3, and Svelte 5 with all states (loading, empty, error, success, idle), TypeScript strict, WCAG 2.2 accessibility, server components (RSC), and…
DB Admin
EliasOulkadi/shokunin
PostgreSQL database administration — backup/restore (pgdump, PITR, WAL archiving), health monitoring (connections, bloat, cache hit ratio, dead tuples), connection pooling (PgBouncer), replication…
DB Sculptor
EliasOulkadi/shokunin
Design database schemas with Prisma/Drizzle, PostgreSQL index strategy (B-tree, GIN, GiST, BRIN, Hash), query optimization (EXPLAIN ANALYZE), migration safety (expand/contract, zero-downtime), and…
Docker
EliasOulkadi/shokunin
Optimize Docker images with multi-stage builds, distroless bases, BuildKit cache mounts, multi-arch builds, compose watch, security hardening (non-root, seccomp, capabilities drop), and…
Error Handler
EliasOulkadi/shokunin
Design error handling, structured logging, and observability with OpenTelemetry (traces, metrics, logs), error classification, recovery patterns (retry with jitter, circuit breaker, bulkhead…
Related
Categories
Questions about Cross Review
What does Cross Review do?
Delegate code review to a subagent running a specific model. Cross Review is an agent skill from EliasOulkadi/shokunin. Delegate code review to a subagent running a specific model.
When should I use Cross Review?
Cross Review fits situations like: explicitly names a model to review changes (review with opus; use sonnet to review; review with gemini); general code review (use code-review skill instead).
How do I install Cross Review in Claude Code?
Run `npx skills add EliasOulkadi/shokunin --skill cross-review -a claude-code`. Or copy the skill folder (.pack/skills/cross-review in EliasOulkadi/shokunin) into .claude/skills/cross-review in your project. Claude Code loads it when a task matches its description.
How do I install Cross Review in Codex?
Run `npx skills add EliasOulkadi/shokunin --skill cross-review -a codex`. Or copy the skill folder (.pack/skills/cross-review in EliasOulkadi/shokunin) into .agents/skills/cross-review in your project. Codex loads it when a task matches its description.
Can I use Cross Review in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add EliasOulkadi/shokunin --skill cross-review -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/cross-review, .gemini/skills/cross-review, .github/skills/cross-review and .opencode/skills/cross-review in your project.
What does Cross Review need to run?
SKILL.md names no scripts, command-line tools or credentials: Cross Review is instructions for the agent only. Compatibility (from SKILL.md): opencode.
Does Cross Review access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Cross Review safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Cross Review use?
Cross Review is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Cross Review use?
About 2.2k tokens (SKILL.md is roughly 8.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Cross Review?
Skills that share tags, products or a category with Cross Review: GitHub Review Iteration (prisma/orm, 48k stars), Cherry Studio PR Review (CherryHQ/cherry-studio, 53k stars), Jevgrep (dzhng/jevgrep, 2.5k stars) and PR Review (jaemk/self_update, 961 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Cross Review?
EliasOulkadi (a GitHub user) maintains it in EliasOulkadi/shokunin, which has 114 GitHub stars. The repository holds 49 skills in this directory. The repository was last updated on October 5, 2026.
Source: EliasOulkadi/shokunin on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.