Review Pending PR Reviews
nrwl/nx
Review, grill, edit, and post pending PR review drafts saved by /review-pr (or its batch/cron runners).
Review agent-native parity for skills, prompts, tools, commands, generated mirrors, repo workflows, and user-facing actions.
Apache-2.0Auto-check: warnings
Install Agent Native Reviewer
The automated check flagged lines worth reading first. See the safety section below.
$ npx skills add udecode/kitcn --skill agent-native-reviewer -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install udecode/kitcn agent-native-reviewer --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/udecode/kitcn.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/agent-native-reviewer .claude/skills/agent-native-reviewer && 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 "agent-native-reviewer" agent skill from https://github.com/udecode/kitcn/tree/main/.agents/skills/agent-native-reviewer into .claude/skills/agent-native-reviewer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agent-native-reviewer", 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/udecode/kitcn/tree/main/.agents/skills/agent-native-reviewerType 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 udecode/kitcn --skill agent-native-reviewer -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install udecode/kitcn agent-native-reviewer --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/udecode/kitcn.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/agent-native-reviewer .agents/skills/agent-native-reviewer && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "agent-native-reviewer" agent skill from https://github.com/udecode/kitcn/tree/main/.agents/skills/agent-native-reviewer into .agents/skills/agent-native-reviewer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agent-native-reviewer", 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 udecode/kitcn --skill agent-native-reviewer -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install udecode/kitcn agent-native-reviewer --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/udecode/kitcn.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/agent-native-reviewer .cursor/skills/agent-native-reviewer && 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 "agent-native-reviewer" agent skill from https://github.com/udecode/kitcn/tree/main/.agents/skills/agent-native-reviewer into .cursor/skills/agent-native-reviewer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agent-native-reviewer", 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/udecode/kitcn.git --path .agents/skills/agent-native-reviewer--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 udecode/kitcn --skill agent-native-reviewer -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install udecode/kitcn agent-native-reviewer --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/udecode/kitcn.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/agent-native-reviewer .gemini/skills/agent-native-reviewer && 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 "agent-native-reviewer" agent skill from https://github.com/udecode/kitcn/tree/main/.agents/skills/agent-native-reviewer into .gemini/skills/agent-native-reviewer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agent-native-reviewer", 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 udecode/kitcn agent-native-reviewerInstalls 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 udecode/kitcn --skill agent-native-reviewer -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/udecode/kitcn.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/agent-native-reviewer .github/skills/agent-native-reviewer && 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 "agent-native-reviewer" agent skill from https://github.com/udecode/kitcn/tree/main/.agents/skills/agent-native-reviewer into .github/skills/agent-native-reviewer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agent-native-reviewer", 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 udecode/kitcn --skill agent-native-reviewer -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install udecode/kitcn agent-native-reviewer --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/udecode/kitcn.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/agent-native-reviewer .opencode/skills/agent-native-reviewer && 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 "agent-native-reviewer" agent skill from https://github.com/udecode/kitcn/tree/main/.agents/skills/agent-native-reviewer into .opencode/skills/agent-native-reviewer/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agent-native-reviewer", 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
agent-native-reviewer- GitHub stars
- 450
- Token cost
- ~2.3k tokens
- SKILL.md length
- 1,189 words
- Files
- 1
- Skills in repo
- 33
- Repo updated
- First seen
- Licence
- Apache-2.0
At a glance
Review agent-native parity for skills, prompts, tools, commands, generated mirrors, repo workflows, and user-facing actions.
- Works in 6 steps: Identify the Changed Surface → Build the Parity Map → Check Source Ownership → …
- SKILL.md covers Use When, Do Not Use When, Core Principles and Codex Capability Ladder, plus 5 more sections
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Agent Native Reviewer is an agent skill from udecode/kitcn. Review agent-native parity for skills, prompts, tools, commands, generated mirrors, repo workflows, and user-facing actions.
Its SKILL.md is about 2.3k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
The repository describes itself as: Convex + Better Auth + tRPC + Drizzle + TanStack Query + shadcn. The licence is Apache-2.0.
Example prompts
- “/agent-native-reviewer”
Workflow steps
6 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 c6010f5. 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 (its code samples are markdown).
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.
Context cost
Agent Native Reviewer loads about 2.3k tokens when it runs. Until then it costs about 37 tokens; SKILL.md has 1,189 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: warnings
The automated check found patterns that need a careful read before installing.
- WarningMentions a credentials file (SSH keys, cloud or package-manager tokens)SKILL.md:64
existing Chrome state, profile, tabs, extensions, cookies, or a site state
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 udecode/kitcn at commit c6010f5, republished under its Apache-2.0 licence (© udecode). 1,189 words, ~2,295 tokens.
Download SKILL.mdSave it as
.claude/skills/agent-native-reviewer/SKILL.md (or your agent's skills folder).- name
- agent-native-reviewer
- description
- Review agent-native parity for skills, prompts, tools, commands, generated mirrors, repo workflows, and user-facing actions.
Agent-Native Reviewer
Review whether an agent can perform, verify, and discover the same meaningful action a user can.
The standard is simple:
user action -> agent route -> source owner -> proof command/artifact -> handoffIf one link is missing, the workflow is not agent-native.
Use When
.agents/**,.claude/**,.codex/**, skills, prompts, hooks, commands, or workflow docs changed.- A feature adds or changes a user-facing action and the repo has agent integration.
- A reviewer asks whether Codex can reproduce, operate, verify, or maintain the same surface a human can.
- A repo is becoming mostly agent-maintained and needs fewer hidden human-only paths.
Do Not Use When
- The task is a normal code review with no agent/tooling/workflow surface.
- The action is intentionally human-only: MFA, OAuth consent, billing payment entry, CAPTCHA, legal acceptance, biometric unlock, or OS permission prompts.
- The change is cosmetic and has no meaningful operation to reproduce.
Core Principles
- Action parity. Every important user action has an agent route.
- Context parity. Agents can see the inputs, state, and constraints needed to act well.
- Source ownership. Agents edit the durable source, not generated mirrors or copied output.
- Proof parity. Agents can verify the outcome with the same authority a maintainer would trust.
- Discoverability. The route is visible from the skill, AGENTS file, command docs, tool schema, or public API docs an agent is expected to read.
- Shared workspace. Agent-created artifacts live where humans can inspect, edit, and commit them.
Codex Capability Ladder
Assume Codex is the agent runtime unless the repo says otherwise. Prefer the most repeatable proof and interaction layer that can cover the action:
- Tests and scripts first. Unit, integration, browser-test, benchmark, typecheck, lint, import smoke, generated-mirror audit, and source audit are the default proof layer. If a behavior can be proved there, do that before driving an app by hand.
- Browser next. Use the Browser plugin for app/browser interaction proof when tests cannot fully prove rendered behavior, native selection, focus, navigation, screenshots, console/network state, or real route behavior.
- Chrome after Browser. Use Chrome when the task depends on the user's existing Chrome state, profile, tabs, extensions, cookies, or a site state that Browser cannot access.
- Computer Use last. Use Computer Use for OS-level or native-app actions that tests, Browser, and Chrome cannot reach. Treat it as powerful but brittle; record exact manual proof and any limits.
Do not jump straight to UI automation when a focused test, command, or source audit gives stronger evidence. Do not claim a behavior is agent-native unless the selected layer can be rerun or described precisely enough for another agent to repeat.
Dotai Integrations
- Use
autogoalfor durable or measurable work. The first checkpoint must copy agent-native requirements into the plan before implementation. - Use
sync-skillswhen a skill, rule, template, or generated mirror crosses repo boundaries. Shared behavior belongs in dotai; repo policy stays local. - Use
sync-visionwhen the missing parity is reusable taste or doctrine, not just one task's mechanics. - Use
resolve-pr-feedbackfor PR review feedback. Do not nest mandatoryautoreview; only recommend it after a complete end-to-end feature when an optional independent second pass could materially help. - Use
hard-cutwhen stale compatibility, fake aliases, dead commands, or duplicate agent routes should be deleted rather than wrapped. - Use
tddordiagnosing-bugswhen the parity gap is a real behavior bug, flaky proof, or unclear failure path.
Review Process
1. Identify the Changed Surface
Classify the action surface:
| Surface | Examples |
|---|---|
| Skill/workflow | SKILL.md, .mdc, plan template, generated mirror, lockfile |
| Command/tool | CLI command, script, MCP tool, GitHub workflow, package script |
| Public API | exported package API, docs example, release artifact |
| Product action | button, form, keyboard shortcut, browser route, OS/device path |
| Maintainer action | issue triage, PR feedback, security advisory, release lane |
For incremental reviews, start from the changed files and expand only to the source owner, generated mirror, lockfile, command, or docs that prove parity.
2. Build the Parity Map
Use this map for every meaningful action:
| User action | Agent route | Source owner | Mirror/lock/doc | Proof | Status |
|---|---|---|---|---|---|
| action | skill/tool/command | file/path | generated/config/doc | command/artifact | pass/gap/N/A |
Status rules:
pass: source owner, route, proof, and discoverability are present.gap: agent cannot safely perform or verify the action.N/A: intentionally human-only or outside the current repo authority, with a concrete reason.
Show full SKILL.md (494 more words)Show less
3. Check Source Ownership
Flag any workflow that asks agents to edit output instead of source.
Common source boundaries:
| Output | Source owner |
|---|---|
.agents/skills/**/SKILL.md installed mirror | external skill package or .agents/rules/** source |
.claude/skills/**/SKILL.md installed mirror | external skill package or .agents/rules/** source |
root AGENTS.md generated block | .agents/AGENTS.md or repo generator input |
| generated barrel/export file | package source plus barrel generator command |
| generated docs/template output | registry/package/docs source named by repo policy |
| copied skill in many repos | dotai source plus Skills CLI install/update |
Finding severity is high when the wrong owner would make future agents lose the fix during sync.
4. Check Agent Route
The route must be usable without hidden human context:
- skill name or tool is discoverable from the available skill list,
AGENTS.md, README, or command docs; - arguments are explicit enough for an agent to call correctly;
- required credentials, browser state, local env, or external access are named;
- authority boundaries are clear: read-only, patch, commit, push, comment, merge, publish, delete.
Do not demand a wrapper skill for every small action. Prefer patching the owner skill, AGENTS routing, or command docs when an existing route fits.
5. Check Proof
Every important agent action needs a verification path:
- exact command, cwd, and expected result;
- generated mirror or lockfile audit;
- browser route/screenshot/console caveat when UI behavior changed;
- issue/PR/security fetch when public maintainer state changed;
- source audit when the claim is structural;
N/Areason when no automated proof fits.
Proof must run in the owning workspace. A downstream install is not proved by a source-package validation alone, and a source-package change is not proved by a downstream command alone.
6. Check Context
Agents need the same operational context a human maintainer uses:
- the relevant vision/doctrine file is linked or routed;
- project-specific commands live in the project, not in dotai;
- final handoff tells the user what changed, what needs attention, and what was not verified;
- recurring misses are written into the durable owner, not left in chat.
If the missing context is reusable taste, route to sync-vision. If it is
workflow mechanics, patch the skill/template. If it is one-off, record it in the
active plan or final handoff.
Findings
Use severity by consequence:
P0: agent can perform a destructive/public action without authority or proof.P1: important user or maintainer action has no safe agent route, or the source/generated boundary is wrong.P2: route exists but is hard to discover, lacks proof, or will rot during sync.P3: wording or docs polish that would improve agent success but is not a blocker.
Suppress low-confidence guesses. If runtime observation is required, ask for or run the proof instead of inventing a finding.
What Not To Flag
- Human-only security/legal/payment ceremonies.
- Cosmetic UI actions without operational meaning.
- Missing automation for rare admin tasks when the repo has no safe authority model for agents.
- A repo-local fork that intentionally keeps product policy out of dotai.
- A generated mirror that is stale only because the required sync command has not run yet; request the sync/proof instead.
Output Format
## Agent-Native Review
### Verdict
PASS | NEEDS WORK
### Capability Map
| User action | Agent route | Source owner | Mirror/lock/doc | Proof | Status |
|---|---|---|---|---|---|
### Findings
1. [P1] Title -- `file:line`
Impact: ...
Fix: ...
Proof: ...
### Accepted / Rejected
- Accepted: ...
- Rejected: ... because ...
### Verification
- command or source audit -> result
### Needs Attention
- ...© udecode, Apache-2.0. 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 .agents/skills/agent-native-reviewer of udecode/kitcn.
Open the folder on GitHubat commit c6010f5
Compare with similar skills
Agent Native Reviewer 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 |
|---|---|---|---|---|---|---|
| Agent Native Reviewer this skilludecode/kitcn | 450 | — | ~2.3k | Automated safety check: Warn | Apache-2.0 | |
| Review Pending PR Reviewsnrwl/nx | 29k | — | ~3.9k | Automated safety check: Pass | MIT | |
| Reviewthedaviddias/Front-End-Checklist | 74k | — | ~556 | Automated safety check: Pass | MIT | |
| ReviewClickHouse/ClickHouse | 50k | — | ~8k | Automated safety check: Notes | Apache-2.0 | |
| Docling Pull Request Reviewdocling-project/docling | 69k | — | ~1k | Automated safety check: Pass | MIT | |
| Pre-Landing PR Reviewgarrytan/gstack | 136k | — | ~19k | Automated safety check: Notes | MIT |
Similar skills
Review
thedaviddias/Front-End-Checklist
A skill your agent uses when applies to product pages, local business pages, recipes, apps, books, and any page that aggregates user reviews.
Review
ClickHouse/ClickHouse
Review a ClickHouse Pull Request for correctness, safety, performance, and compliance.
Docling Pull Request Review
docling-project/docling
Reviews or re-reviews a Docling pull request in fixed stages, with findings that can be reproduced and an explicit record of every check that was run.
Pre-Landing PR Review
garrytan/gstack
Reviews your diff against the base branch before merge, looking for SQL safety problems, LLM trust boundary violations, conditional side effects and other structural issues.
- Official
Docs Parity Reviewer
pydantic/pydantic-ai
Use as the final documentation gate before a pydantic-ai-harness capability PR merges.
More from udecode/kitcn
All 33 skills in this repoWalkthrough
udecode/kitcn
Create a short annotated visual walkthrough from real final-state screenshots or rendered artifacts.
Avoid Feature Creep
udecode/kitcn
Prevent feature creep when building software, apps, and AI-powered products.
Changeset Resolve
udecode/kitcn
Repair an unreleased .changeset/.md file so it matches the real branch delta against main.
Convex Release Audit
udecode/kitcn
Audit newer Convex npm releases against kitcn. An agent skill from udecode/kitcn.
Jotai X
udecode/kitcn
A skill your agent uses when working with Jotai X stores (createAtomStore), accessing state in components or callbacks, persisting state to cookies or localStorage
Linear Backlog
udecode/kitcn
Run a scoped Linear backlog autonomously as a sequence of maximal safe parallel batches by composing orchestrator, autogoal, and task.
Questions about Agent Native Reviewer
What does Agent Native Reviewer do?
Review agent-native parity for skills, prompts, tools, commands, generated mirrors, repo workflows, and user-facing actions. Agent Native Reviewer is an agent skill from udecode/kitcn. Review agent-native parity for skills, prompts, tools, commands, generated mirrors, repo workflows, and user-facing actions.
How do I install Agent Native Reviewer in Claude Code?
Run `npx skills add udecode/kitcn --skill agent-native-reviewer -a claude-code`. Or copy the skill folder (.agents/skills/agent-native-reviewer in udecode/kitcn) into .claude/skills/agent-native-reviewer in your project. Claude Code loads it when a task matches its description.
How do I install Agent Native Reviewer in Codex?
Run `npx skills add udecode/kitcn --skill agent-native-reviewer -a codex`. Or copy the skill folder (.agents/skills/agent-native-reviewer in udecode/kitcn) into .agents/skills/agent-native-reviewer in your project. Codex loads it when a task matches its description.
Can I use Agent Native Reviewer 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 udecode/kitcn --skill agent-native-reviewer -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/agent-native-reviewer, .gemini/skills/agent-native-reviewer, .github/skills/agent-native-reviewer and .opencode/skills/agent-native-reviewer in your project.
What does Agent Native Reviewer need to run?
SKILL.md names no scripts, command-line tools or credentials: Agent Native Reviewer is instructions for the agent only.
Does Agent Native Reviewer 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 Agent Native Reviewer safe to install?
Our automated static check of SKILL.md flagged 1 warning(s): mentions a credentials file (ssh keys, cloud or package-manager tokens). Read the flagged lines before installing; the check is not a guarantee either way.
What licence does Agent Native Reviewer use?
Agent Native Reviewer is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Agent Native Reviewer use?
About 2.3k tokens (SKILL.md is roughly 9.2k 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 Agent Native Reviewer?
Skills that share tags, products or a category with Agent Native Reviewer: Review Pending PR Reviews (nrwl/nx, 29k stars), Review (thedaviddias/Front-End-Checklist, 74k stars), Review (ClickHouse/ClickHouse, 50k stars) and Docling Pull Request Review (docling-project/docling, 69k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Agent Native Reviewer?
udecode (a GitHub organization) maintains it in udecode/kitcn, which has 450 GitHub stars. The repository holds 33 skills in this directory. The repository was last updated on October 1, 2026.
Source: udecode/kitcn on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.