guard-skills
Free open-source guard skills that catch hallucinated APIs and assertion-free tests in AI-generated code before your agent commits.
If an AI agent writes any of your code, Guard Skills is a low-ceremony way to intercept the errors that look fine on first read — hallucinated APIs and assertion-free tests. At 16.1K total installs it has enough adoption to trust the maintenance, and every guard is inspectable on GitHub. Start with clean-code-guard and test-guard; the WordPress pair only earns its slot if that's your stack. It won't replace ESLint or SonarQube and it won't replace human review, but it catches a class of error those tools were never designed for.
Verified 8d ago · liveness 72/100 · cite: rightaichoice.com/tools/guard-skills
- Developers using AI coding agents who want to catch hallucinated APIs before commit
- Teams shipping AI-generated code that needs guardrails against assertion-free ghost tests
- WordPress and WooCommerce developers who want platform-specific AI error prevention
- Engineering leads adding quality gates to existing agent workflows without buying a platform
- Non-developers or anyone not running an AI coding agent in their workflow
- Teams whose review process already catches hallucinated APIs and weak tests reliably
- Projects needing runtime monitoring, performance profiling, or production observability
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Guard Skills if your code review already reliably catches hallucinated APIs and assertion-free tests, or if you don't run an AI coding agent in your workflow at all.
There's no licence fee, but the guards run inside your agent, so you pay the token cost of the extra inspection passes on every commit they gate.
Free and open source, with no paid tier and no usage caps — the entire cost is the agent compute the guards consume when they run. That puts it well below commercial quality-gate platforms such as SonarQube or Codacy, which charge per line of code or per seat. If your budget is zero and your need is narrow (AI-specific defects, not general static analysis), it fits; if you need dashboards, SLAs, or custom rule authoring hosted for you, a paid platform is the fit.
In short
guard-skills — Free open-source guard skills that catch hallucinated APIs and assertion-free tests in AI-generated code before your agent commits. Best for Developers using AI coding agents who want to catch hallucinated APIs before commit, Teams shipping AI-generated code that needs guardrails against assertion-free ghost tests, WordPress and WooCommerce developers who want platform-specific AI error prevention. Free to use.
What's new in guard-skills
Checked 8 days agoAcross the latest 1 update: 1 changelog entry.
What people actually say about guard-skills — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
23 mentions across 3 sources (Hacker News, YouTube, GitHub) · researched Aug 30, 2026.
Average across the 3 sources that answered — each source counts once, not each post.
- +15+ quality gates catching hallucinated APIs, ghost tests, and verbose comments.
- +Free forever and fully open-source, allowing inspectability and customization.
- +Seamless integration with 10+ AI agents like Claude, Cursor, and Codex.
- +Proactive at commit time, catching issues before they hit production.
- +Specialized WordPress and WooCommerce guards target ecosystem-specific mistakes.
- −Occasional enforcement gaps reduce effectiveness (guard runs at 40-60% efficiency).
- −Threshold configuration requires editing code, not user-friendly settings.
- −No dedicated support; relies on GitHub issues for help.
- −Limited community feedback beyond HN and GitHub—buzz is still niche.
- −Not a replacement for comprehensive static analysis; narrow focus.
- • Time spent customizing thresholds to get consistent behavior.
Viability Score
How well maintained and how widely used is guard-skills? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- clean-code-guard: detects hallucinated and non-existent APIs in AI-generated code
- test-guard: validates that generated tests actually assert meaningful behavior
- docs-guard: keeps documentation accurate and consistent with the code
- wp-guard: catches WordPress-specific mistakes common in AI output
- woo-guard: handles WooCommerce extension issues from AI agents
- Install the full pack with npx skills add amelnagdy/guard-skills
- Add or remove individual guard skills independently
- Works with Claude Code, Cursor, Codex, and GitHub Copilot
- Works with Windsurf, Gemini, Cline, AMP, Antigravity, and OpenClaw
- Open source guard logic on GitHub — read, tweak, and contribute
- Quality gates run at commit or integration time inside the agent
- Routine security audits by Vercel on skills.sh
- Free with no paid tier or usage limit
- Distributed via the open-source skills CLI (github.com/vercel-labs/skills)
About guard-skills
Guard Skills is a free, open-source pack of five quality-gate skills for AI coding agents. It installs into the agents already in your workflow — Claude Code, Cursor, Codex, GitHub Copilot, Windsurf, Gemini, Cline, AMP, Antigravity, or OpenClaw — and inspects what the agent produces, flagging the failure modes LLMs repeat: invented API calls, tests that assert nothing, docs that drift from the code, and platform-specific mistakes in WordPress and WooCommerce work. Install is a single command, `npx skills add amelnagdy/guard-skills`, and the pack ships as five separately installable skills: clean-code-guard (4.2K installs) for hallucinated and non-existent APIs, test-guard (3.8K) for tests that are real and meaningful, docs-guard (3.6K) for documentation accuracy, wp-guard (2.2K) for WordPress-specific AI errors, and woo-guard (2.2K) for WooCommerce extension issues. Combined installs sit at 16.1K, up from the 14.8K in the seed data and the 11.3K recorded in an earlier changelog entry. Each skill is open source on GitHub, so you can read the guard logic, adjust thresholds, and contribute fixes. The distinction from general linters like ESLint or SonarQube is the target: those check syntax and patterns, these guards check the categories of error AI agents generate specifically. It sits alongside your linter rather than replacing it.
Behind the Verdict
Guard Skills solves a narrow, real problem: AI coding agents produce a specific signature of defect that conventional tooling misses. A hallucinated function that doesn't exist in the installed version of a library often looks plausible and passes a linter because the linter never checks whether the API is real. A generated test that runs green but asserts nothing passes CI. Docs that describe a function the agent later renamed drift silently. Guard Skills targets exactly these categories. The pack is five skills, each installed independently. clean-code-guard (4.2K installs) is the headline: it detects hallucinated and non-existent APIs. test-guard (3.8K) validates that generated tests actually assert meaningful behavior rather than exercising code and calling it a day. docs-guard (3.6K) keeps documentation consistent with the code. wp-guard (2.2K) and woo-guard (2.2K) handle WordPress and WooCommerce conventions, where AI agents routinely get platform-specific details wrong. Where it fits: you already run an AI coding agent and you've been burned by a plausible-looking commit that broke at runtime. Installing one skill takes a single command and the guards run as quality gates inside the agent at commit or integration time. The logic is open source, so if the thresholds are wrong for your codebase you can change them — this is a meaningful advantage over closed quality platforms. Where it doesn't: the guards are rule-based and will not catch every AI-specific issue, and they sit alongside human review rather than replacing it. There is no dashboard, no API, and no custom rule creation outside modifying the open source code. If your review process already catches hallucinated APIs and weak tests reliably, adding another gate buys you little. And if you need runtime monitoring, performance profiling, or production observability, this is not that tool. The WordPress pair only makes sense if WordPress is your stack — otherwise it's two skills you don't need.
Researching guard-skills? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas guard-skills actually fits — and what changes day-one when you adopt it.
Adds clean-code-guard and test-guard with `npx skills add amelnagdy/guard-skills`, then lets the agent write a feature and its tests. Before commit, clean-code-guard flags a library function the agent invented and test-guard flags a test that runs but asserts nothing.
Outcome: Both defects get fixed in the same session instead of surfacing at runtime or during review, at no licence cost.
Installs wp-guard and woo-guard alongside the general guards, then has the agent scaffold a WooCommerce extension. The guards flag WordPress-specific conventions the agent got wrong and the WooCommerce API misuse.
Outcome: The plugin ships without the platform-specific mistakes that would otherwise fail in staging.
Standardises on the pack across a team already using Cursor and GitHub Copilot, keeping the guard logic in the repo so thresholds can be reviewed like any other code.
Outcome: A shared quality gate for AI-generated code that costs nothing to licence and is auditable in version control.
Use Cases
- Automatically check AI-generated code for hallucinated function names and APIs before commit.
- Ensure AI-written unit tests actually run and assert real behaviors instead of passing empty.
- Catch documentation that describes features or signatures not present in the code.
- Apply WordPress and WooCommerce guards when building plugins or themes with an AI agent.
Limitations
- The guards are rule-based and may not catch all AI-specific issues; they complement but do not replace human review.
- Installation requires the skills CLI and a compatible AI agent.
- No API, dashboard, or custom rule creation is available outside modifying the open source code.
- As the skills.sh documentation notes, 'we cannot guarantee the quality or security of every skill listed on skills.sh.' It sits alongside ESLint or SonarQube rather than replacing them — those check syntax and patterns, these check AI-specific error categories.
as of 2026-09-30
Verification history
We have re-verified guard-skills 9 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 9 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published guard-skills tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0
Ideal for
Any developer or team running an AI coding agent who wants AI-specific defect gates without a licence fee or sales cycle.
What this tier adds
Starting tier — all five guard skills included, open source, no paid tier and no usage caps.
Where the pricing makes sense
The company stage and team size where guard-skills's pricing actually pencils out — and where peers do it cheaper.
Free and open source, with no paid tier and no usage caps — the entire cost is the agent compute the guards consume when they run. That puts it well below commercial quality-gate platforms such as SonarQube or Codacy, which charge per line of code or per seat. If your budget is zero and your need is narrow (AI-specific defects, not general static analysis), it fits; if you need dashboards, SLAs, or custom rule authoring hosted for you, a paid platform is the fit.
Setup time & first value
How long it actually takes to get something useful out of guard-skills — broken out by persona, not the marketing-page minute.
For a solo developer already running Claude Code, Cursor, or Copilot: one command (`npx skills add amelnagdy/guard-skills`) and you're gated on the next commit — under five minutes. For a team standardising the pack across multiple agents and tuning thresholds by editing the open source guard logic, budget an hour or two to agree on which skills to install and review the defaults.
Switching to or from guard-skills
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From manual review only: install clean-code-guard and test-guard via the skills CLI and let them gate commits before the review step.
- ↗To a commercial quality platform: if you need dashboards, hosted custom rules, or an SLA, move to a paid static-analysis or quality-gate service and keep Guard Skills only for the AI-specific checks it does uniquely.
Integrations
Resources & Guides
Tutorials & Learning
PATH of EXILE: GUARD SKILLS - Which Should You Use in Your Build?
ZiggyD Gaming
Your Guide to MUST-HAVE Guard Skills!
Hoop Doctors
YouTube returned 6 videos for “guard-skills”, and we withheld 4: 4 did not mention guard-skills. Showing the 2 we can prove are about guard-skills.
Official links
Tools that pair well with guard-skills
Common stack mates teams adopt alongside guard-skills, with the specific reason each pairing earns its keep.
Agents Shipgate
Open-source CLI and GitHub Action that returns a deterministic merge verdict on AI-generated agent PRs before the code lands.
DevoxxGenieIDEAPlugin
Free, open-source AI code assistant plugin for JetBrains IDEs with spec-driven development and agent mode
Continue
Open-source AI coding agent for VS Code and JetBrains, acquired by Cursor in January 2026 and now an unmaintained codebase you fork, not subscribe to.
Featured Head-to-Head Comparisons
Guard Skills vs Bito
Choose Bito if your team operates across multiple repos and needs deep architectural awareness for AI coding agents, with features like cross-repo impact analysis and automated design docs. Choose Guard Skills if you want a free, open-source safety net to catch common AI mistakes like hallucinated APIs or weak tests, especially for WordPress/WooCommerce projects. They solve different problems — Bito provides system-wide context, Guard Skills provides lightweight quality checks — and can be complementary.
Guard Skills vs Cognition Ai
Choose Cognition AI (Devin) if you're an enterprise team needing an autonomous engineer that can handle multi-step tasks like bug triage, legacy modernization, and cross-platform builds—backed by a financial guarantee. Choose Guard Skills if you're an individual developer or small team using AI coding agents and want free, open-source quality gates to catch common AI failures quickly. They serve different layers: Devin is the doer, Guard Skills is the checker.
Guard Skills vs Poolside Ai
If you're building high-stakes software in finance or defense and need custom models deployed inside a VPC with enterprise governance, Poolside AI is the only option. But for most teams using AI coding agents today, guard-skills is a no-brainer: free, open-source, and instantly catches AI-specific mistakes in code, tests, and docs. Start with guard-skills; graduate to Poolside when compliance demands it.
Guard Skills vs Replit Agent
If you need to rapidly build and deploy full-stack apps from natural language, Replit Agent is the clear choice with its integrated IDE and one-click deployment. But if you're already using AI coding agents and want to catch common failures like hallucinated APIs or fake tests, Guard Skills is a no-brainer free addition. They're complementary: use Replit as the builder, Guard Skills as the checker.
Guard Skills vs Mastra
Mastra is the right choice if you're building complex, production-grade AI agent systems with multi-step workflows, durable execution, and robust observability. Guard-skills is ideal if your primary concern is ensuring quality of AI-generated code, especially in WordPress/WooCommerce. They serve different needs and can even be complementary.
Guard Skills vs Marvin
If you're a Python developer wanting to embed LLM logic into your code with decorators for extraction, classification, or agents, Marvin is the right choice. If you use AI coding agents and need automated quality gates to catch hallucinated APIs or fake tests, guard-skills fills that specific gap. Both are free and open-source, so cost isn't a differentiator—your workflow determines the pick.
Alternatives to guard-skills
View all
Agents Shipgate
Open-source CLI and GitHub Action that returns a deterministic merge verdict on AI-generated agent PRs before the code lands.
DevoxxGenieIDEAPlugin
Free, open-source AI code assistant plugin for JetBrains IDEs with spec-driven development and agent mode
Frequently Asked Questions
Categories
Best-of guides
Used guard-skills? Help shape our editorial sentiment research.