Agent skill

Plugin Review

by base in base/skills

Validate and review Base MCP plugin files against the Plugin Specification.

MITAuto-check passedAgent Workflows

Install Plugin Review

skills CLI
$ npx skills add base/skills --skill plugin-review -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install base/skills plugin-review --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/base/skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/plugin-review .claude/skills/plugin-review && rm -rf skills-src

Use ~/.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/

Facts

Skill name
plugin-review
GitHub stars
121
Token cost
~1.3k tokens
SKILL.md length
543 words
Files
5 (incl. references)
Skills in repo
4
Repo updated
First seen
Licence
MIT

At a glance

Validate and review Base MCP plugin files against the Plugin Specification.

  • Works in 5 steps: Fetch the current spec (it changes —… → Read the plugin file in full. If… → Static conformance evaluation — assess… → …
  • Writing a new plugin
  • SKILL.md covers Workflow, Multiple PRs and Critical gotchas
  • Calls gh and curl

What it does

Plugin Review is an agent skill from base/skills. Validate and review Base MCP plugin files against the Plugin Specification. Use when writing a new plugin, preparing a plugin PR for submission, self-checking a plugin before opening a PR, or reviewing someone else's plugin PR. Triggers on requests like "check my plugin against the spec", "validate this plugin file", "review my plugin PR", "does this plugin conform to the spec", "prepare my plugin for submission", "write a plugin for X protocol".

Its SKILL.md is about 1.3k tokens, which your agent loads only when the skill is triggered. The skill folder holds 5 other files, including reference files (for example `references/comment-guidelines.md`, `references/evaluation-criteria.md` and `references/live-testing.md`).

It sits in Agent Workflows. It works with Model Context Protocol. The repository describes itself as: npx skills add base/skills. The licence is MIT.

When your agent uses it

  • Writing a new plugin
  • Preparing a plugin PR for submission
  • Self-checking a plugin before opening a PR
  • Reviewing someone elses plugin PR

Example prompts

  • “s plugin PR. Triggers on requests like”
  • “validate this plugin file”
  • “review my plugin PR”
  • “/plugin-review”

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

  1. Fetch the current spec (it changes — never rely on a stale copy)
  2. Read the plugin file in full. If reviewing a PR
  3. Static conformance evaluation — assess against every dimension in references/evaluation-criteria.md (includes compliance/language checks…
  4. (Optional) Live API / SDK verification — exercise the documented endpoints/SDK/contracts with read-only calls. See…
  5. Save the report. If reviewing a PR and asked to comment, draft a PR comment from the report using references/comment-guidelines.md and…

What it can do on your machine

Read from SKILL.md and the folder at commit 4f3d8f3. 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

    Shell commands in SKILL.md call:

    • gh
    • curl

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use gh and curl, which can reach the network depending on how they are called.

    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

Plugin Review loads about 1.3k tokens when it runs, and up to ~6.3k if it reads all its reference files. Until then it costs about 116 tokens; SKILL.md has 543 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~116
When it runs · the whole SKILL.md, loaded when a task matches
~1.3k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~6.3k

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 base/skills at commit 4f3d8f3, republished under its MIT licence (© base). 543 words, ~1,269 tokens.

Download SKILL.mdSave it as .claude/skills/plugin-review/SKILL.md (or your agent's skills folder). This skill also uses 4 other files; get the full folder from GitHub.
name
plugin-review
description
Validate and review Base MCP plugin files against the Plugin Specification. Use when writing a new plugin, preparing a plugin PR for submission, self-checking a plugin before opening a PR, or reviewing someone else's plugin PR. Triggers on requests like "check my plugin against the spec", "validate this plugin file", "review my plugin PR", "does this plugin conform to the spec", "prepare my plugin for submission", "write a plugin for X protocol".
license
Complete terms in LICENSE.txt

Plugin Review

Validate Base MCP plugin files against the current Plugin Specification. Produces a conformance report with actionable findings.

Works for both authors (self-check before submitting a PR) and reviewers (evaluate an incoming PR).

Workflow

  1. Fetch the current spec (it changes — never rely on a stale copy):

    bash
    curl -s https://raw.githubusercontent.com/base/skills/master/skills/base-mcp/references/plugin-spec.md

    Related docs worth reading: references/custom-plugins.md, references/approval-mode.md, references/batch-calls.md, and SKILL.md (the root skill). Existing native plugins under skills/base-mcp/plugins/ are the precedent for conventions.

  2. Read the plugin file in full. If reviewing a PR:

    bash
    gh pr view <n> --repo base/skills --json title,body,number,headRefName,files,additions,deletions
    gh pr diff <n> --repo base/skills
    # raw plugin file (diff may be truncated):
    gh api "repos/base/skills/pulls/<n>/files" --jq '.[] | select(.filename|endswith(".md")) | .raw_url'
  3. Static conformance evaluation — assess against every dimension in references/evaluation-criteria.md (includes compliance/language checks and high-risk category gates). Write the report using references/report-template.md.

  4. (Optional) Live API / SDK verification — exercise the documented endpoints/SDK/contracts with read-only calls. See references/live-testing.md. For perps/prediction-market/gambling plugins, also run the geoblock verification (compare frontend vs API access restrictions). Append a ## Live API / SDK Verification section to the report. This routinely overturns doc claims (fabricated/locked endpoints, broken hosts, wrong response shapes).

  5. Save the report. If reviewing a PR and asked to comment, draft a PR comment from the report using references/comment-guidelines.md and post with: gh pr comment <n> --repo base/skills --body-file <comment-file>.

Multiple PRs

Evaluate PRs in parallel by spinning up one sub-agent per PR (each writes its own report + returns a short verdict summary). Hand each sub-agent: the spec (or its raw URL), the PR number, the evaluation criteria, the report template, and the comment guidelines.

Show full SKILL.md (312 more words)Show less

Critical gotchas

These recur and are easy to get wrong — full detail in references/evaluation-criteria.md:

  • Smart-account signatures break naive signing. The default Base MCP wallet is a smart contract; sign returns a variable-length ERC-1271/6492 signature (>200 bytes), not a 65-byte EOA sig. Any plugin that splices a signature into a fixed-width calldata slot, or bakes an off-chain EIP-712 signature into calldata (e.g. Permit2 buildCallWithPermit2), is broken for that wallet. Correct pattern: onchain allowance grants.
  • irreversible risk is NOT for every onchain write. The spec says "flag when worth emphasizing." Pure swaps use slippage (precedent: Uniswap, Aerodrome carry [slippage], not irreversible). Reserve irreversible for asymmetric/severe cases (perps/liquidation, token launches/rug). Do not demand it on swaps.
  • Don't self-register. A plugin PR must NOT edit the SKILL.md plugins table, the Integration Types "Examples" cell, or the "Existing Plugin Conformance" table — those are maintainer-managed (codified in plugin-spec.md "Contribution Scope"). The only sanctioned shared-file edit is appending a genuinely net-new tag to the vocabulary list. Limit the diff to plugins/<slug>.md (+ that tag line).
  • version is the plugin-doc version — not the npm/package version and not a global spec version. The spec mandates no specific starting number.
  • Verify claims, don't trust them. Auth models, allowlist completeness, response shapes, and contract addresses are frequently wrong in the doc. Probe them (live testing).
  • Reference links from a plugin file must use ../references/... (plugin files live in plugins/, refs in references/).
  • Neutral language is mandatory. No yield/rate/performance claims, no "you should buy X", no "always deposit here", no defaulting to specific tokens. Steering language is a blocker.
  • Perps, prediction markets, and privacy plugins need legal review before inclusion as native plugins. Flag this as a pre-merge process gate in the report.
  • API geoblock parity. If the protocol's frontend geoblocks US IPs (or others), the API must enforce equivalent restrictions. If it doesn't, Base MCP risks being a circumvention tool — flag as a blocker.

© base, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 4 other files (references) in .claude/skills/plugin-review of base/skills.

  • SKILL.md
  • references/comment-guidelines.md
  • references/evaluation-criteria.md
  • references/live-testing.md
  • references/report-template.md

Open the folder on GitHubat commit 4f3d8f3

Compare with similar skills

Plugin 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.

Plugin Review compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Plugin Review this skillbase/skills121—~1.3kAutomated safety check: PassMIT
MCP Server Builderanthropics/skills180k62 repos~2.3kAutomated safety check: PassApache-2.0
MCP Server BuildershareAI-lab/learn-claude-code78k5 repos~1.2kAutomated safety check: PassMIT
MCP Integration for Pluginsanthropics/claude-plugins-official37k11 repos~3.1kAutomated safety check: PassApache-2.0
Fastmcp Client CLIPrefectHQ/fastmcp28k1 repos~823Automated safety check: PassApache-2.0
MemPalace Memory SearchMemPalace/mempalace59k—~1.4kAutomated safety check: PassMIT

Similar skills

  • MCP Server Builder

    anthropics/skills

    Official

    Guides the design and implementation of Model Context Protocol servers in TypeScript or Python, from tool naming and error messages to evaluation.

    180k GitHub starsUsed in 62 repos~2.3k tokens
    Agent WorkflowsAuto-check passed
  • MCP Server Builder

    shareAI-lab/learn-claude-code

    Walks through building MCP servers in Python or TypeScript that expose tools, resources and prompts to Claude, with templates, registration and testing.

    78k GitHub starsUsed in 5 repos~1.2k tokens
    Agent WorkflowsAuto-check passed
  • MCP Integration for Plugins

    anthropics/claude-plugins-official

    Official

    Explains how to bundle Model Context Protocol servers in a Claude Code plugin, covering config files, stdio, SSE, HTTP and WebSocket server types, and authentication.

    37k GitHub starsUsed in 11 repos~3.1k tokens
    Agent WorkflowsAuto-check passed
  • Fastmcp Client CLI

    PrefectHQ/fastmcp

    Query and invoke tools on MCP servers using fastmcp list and fastmcp call.

    28k GitHub starsUsed in 1 repo~823 tokens
    Agent WorkflowsAuto-check passed
  • MemPalace Memory Search

    MemPalace/mempalace

    Mines project files and conversation exports into a local, searchable memory palace and recalls past work by semantic search through the mempalace CLI.

    59k GitHub stars~1.4k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Crush Configuration

    charmbracelet/crush

    Explains how to configure the Crush coding agent with crushrc or crush.json, covering providers, models, LSPs, MCP servers, hooks, permissions and config precedence.

    29k GitHub stars~3.7k tokensUpdated today
    Agent WorkflowsAuto-check passed

More from base/skills

  • Vibenet

    base/skills

    Build on vibenet — Base's devnet for native account abstraction (EIP-8130), payer gas sponsorship (ERC-8168), and Cobalt 200ms native blocks.

    121 GitHub stars~2.7k tokensUpdated 5 days ago
    Auto-check passed
  • Build On Base

    base/skills

    Complete Base development playbook. An agent skill from base/skills.

    121 GitHub stars~804 tokensUpdated 5 days ago
    Auto-check passed
  • Base MCP

    base/skills

    Base MCP — gives your AI assistant access to a Base Account via the Base MCP server (wallet-mcp.coinbase.com).

    121 GitHub stars~3.5k tokensUpdated 5 days ago
    Auto-check passed

Categories

Questions about Plugin Review

What does Plugin Review do?

Validate and review Base MCP plugin files against the Plugin Specification. Plugin Review is an agent skill from base/skills. Validate and review Base MCP plugin files against the Plugin Specification.

When should I use Plugin Review?

Plugin Review fits situations like: writing a new plugin; preparing a plugin PR for submission; self-checking a plugin before opening a PR; reviewing someone elses plugin PR.

How do I install Plugin Review in Claude Code?

Run `npx skills add base/skills --skill plugin-review -a claude-code`. Or copy the skill folder (.claude/skills/plugin-review in base/skills) into .claude/skills/plugin-review in your project. Claude Code loads it when a task matches its description.

How do I install Plugin Review in Codex?

Run `npx skills add base/skills --skill plugin-review -a codex`. Or copy the skill folder (.claude/skills/plugin-review in base/skills) into .agents/skills/plugin-review in your project. Codex loads it when a task matches its description.

Can I use Plugin 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 base/skills --skill plugin-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/plugin-review, .gemini/skills/plugin-review, .github/skills/plugin-review and .opencode/skills/plugin-review in your project.

What does Plugin Review need to run?

Going by SKILL.md and its folder, Plugin Review needs the command-line tools its instructions call (gh and curl).

Does Plugin Review access the network?

SKILL.md contains no URLs. Its commands use gh and curl, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Plugin 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 Plugin Review use?

Plugin Review is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Plugin Review use?

About 1.3k tokens (SKILL.md is roughly 5.1k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 5k tokens, read only when the agent opens those files.

What are the alternatives to Plugin Review?

Skills that share tags, products or a category with Plugin Review: MCP Server Builder (anthropics/skills, 180k stars), MCP Server Builder (shareAI-lab/learn-claude-code, 78k stars), MCP Integration for Plugins (anthropics/claude-plugins-official, 37k stars) and Fastmcp Client CLI (PrefectHQ/fastmcp, 28k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Plugin Review?

base (a GitHub organization) maintains it in base/skills, which has 121 GitHub stars. The repository holds 4 skills in this directory. The repository was last updated on October 1, 2026.

Source: base/skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.