← Back to Skills Marketplace
anderskev

Review Elixir

by Kevin Anderson · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ⚠ suspicious
116
Downloads
0
Stars
1
Active Installs
2
Versions
Install in OpenClaw
/install review-elixir
Description
Comprehensive Elixir/Phoenix code review with optional parallel agents
Usage Guidance
This skill generally does what it says: it runs repository checks and linters and consolidates findings. Before installing or running it: 1) Confirm your runtime has git, Elixir (mix), and grep (and dialyzer/credo if you want those checks); the registry metadata currently omits these required binaries. 2) Inspect the referenced 'beagle-elixir:*' skills (and the 'beagle-elixir:review-verification-protocol') before allowing the skill to load them — they will expand what the agent can do. 3) Be aware the skill reads repository files and will re-open file:line locations when reporting Critical/Major issues — avoid running it on repos with secrets you don't want exposed. 4) If you need higher assurance, ask the publisher for a source/homepage and an updated manifest that declares required binaries and describes the other skills it loads. My confidence is medium because the main issues are an apparent manifest omission and the delegation to other, unreviewed skills; these could be benign oversights but should be verified.
Capability Analysis
Type: OpenClaw Skill Name: review-elixir Version: 1.0.1 The skill bundle provides a legitimate and well-structured workflow for performing Elixir and Phoenix code reviews. It utilizes standard development tools such as `mix format`, `credo`, and `dialyzer` to verify code quality and includes logic to detect specific technologies like LiveView and Oban. The instructions in SKILL.md are focused on ensuring review accuracy through 'hard gates' and verification steps, with no evidence of malicious intent, data exfiltration, or unauthorized system access.
Capability Assessment
Purpose & Capability
The SKILL.md explicitly runs git, mix, grep and may call mix credo/dialyzer; yet the registry Requirements list no required binaries. A code-review skill legitimately needs repo and build-tool access (git, Elixir's mix, grep/grep-like tools). The manifest should declare these binaries. Also the skill expects to load many other 'beagle-elixir:*' skills — the manifest doesn't identify or justify those dependencies or provide a homepage/source for vetting.
Instruction Scope
Instructions are narrowly focused on reviewing repository files, running project linters/formatters, and re-reading specific file:line locations — actions that match the stated purpose. The skill instructs loading other review skills and spawning subagents; that expands runtime scope and depends on those skills' behavior, so their contents should be reviewed before trusting end-to-end behavior.
Install Mechanism
This is an instruction-only skill with no install spec or code files, so it does not write or execute bundled code on disk. That lowers installation risk.
Credentials
The skill declares no required environment variables or credentials, which aligns with its purpose. However, it will invoke tooling (git, mix, possibly dialyzer) that may in practice require local configuration or VCS credentials; the manifest should list required binaries and any configuration expectations.
Persistence & Privilege
The skill is not always-on and does not request elevated platform privileges. disable-model-invocation is true, so it cannot autonomously call the model; it does instruct spawning subagents via the agent's Task/Skill tools, which is expected for a parallel review flow.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install review-elixir
  3. After installation, invoke the skill by name or use /review-elixir
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
**Adds strict pre-issue gates for scope, linter, protocol, and evidence to ensure review quality and reproducibility.** - Introduces four "Hard gates" (scope, linter, protocol, evidence) that must be passed in order before reporting any issues. - Clarifies that an explicit file list must be printed or declared empty before proceeding; fabricated issue reports are disallowed. - Requires explicit linter (formatter, Credo, Dialyzer) run status to be reported, and never flags linter-passing style lines as issues. - Mandates that all Critical/Major issues be cross-checked against the verification protocol and code context at FILE:LINE. - Strengthens verification steps before and after reviews to prevent speculative or unsubstantiated reports.
v1.0.0
- Initial release of review-elixir: a comprehensive Elixir/Phoenix code review skill. - Supports sequential and parallel review of Elixir, Phoenix, LiveView, performance, security, and test code, with technology detection. - Strict linter verification: linters are authoritative for style; only flag non-linter issues. - Enforces single-pass review policy: report ALL categories of issues up front, including anticipated downstream problems. - Provides structured output format, explicit severity levels (Critical/Major/Minor/Informational), and review scope rules. - Re-reviews are verification-only; no new findings allowed beyond previous issues.
Metadata
Slug review-elixir
Version 1.0.1
License MIT-0
All-time Installs 1
Active Installs 1
Total Versions 2
Frequently Asked Questions

What is Review Elixir?

Comprehensive Elixir/Phoenix code review with optional parallel agents. It is an AI Agent Skill for Claude Code / OpenClaw, with 116 downloads so far.

How do I install Review Elixir?

Run "/install review-elixir" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Review Elixir free?

Yes, Review Elixir is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Review Elixir support?

Review Elixir is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Review Elixir?

It is built and maintained by Kevin Anderson (@anderskev); the current version is v1.0.1.

💬 Comments