CI Security Scanning With Strix Skill: Add an AI Pentest to Every Pull Request

Has your agent set up Strix in CI so every pull request gets a diff-scoped AI security scan that can block the merge.

Installs: 16.9K; on skills.sh
Stars: 73.2K; on GitHub
Updated: Oct 2026; last repo commit

You ask for security scanning, SAST or a pre-merge security gate, and the agent writes the pipeline for you. It offers two routes. Option A runs the open-source Strix CLI in your own runner, needs Docker and your own LLM key, and keeps scans inside your infrastructure. Option B uses the managed app.strix.ai platform, with no runner, no Docker and no LLM key.

The agent writes a workflow file such as .github/workflows/security.yml, and tells you which repository secrets to add. It does not create the secret values itself. Strix is a separate product, and managed scans use its credits.

Good for

Developers who want a security gate on pull requests
DevOps engineers wiring GitHub Actions or GitLab CI
Teams that need scans to stay inside their own runners

Install ci-security-scanning-with-strix

Claude Code

Copy this prompt.

Install the ci-security-scanning-with-strix skill in Claude Code. Run this command:

npx skills add https://github.com/usestrix/strix --skill ci-security-scanning-with-strix

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js.

From its listing on skills.sh. Source.

What's in the skill

When it loads

Add security scanning to CI/CD with Strix — GitHub Actions, GitLab CI, or any pipeline — so every pull request gets a diff-scoped AI pentest that blocks vulnerable code before it merges, with results as PR comments and SARIF uploaded to code scanning. Covers both the self-hosted open-source CLI (runs in your runner) and the managed app.strix.ai platform (GitHub/GitLab app or API, no runner infra). Use when the user asks to add security scanning, SAST/DAST, pentesting, vulnerability checks, or automated security review to their CI pipeline, pre-merge gate, or PR workflow.

Sections in SKILL.md

GitHub Actions
Optional: upload findings to GitHub code scanning
Other CI systems

Key facts

Author: usestrix
License: Apache-2.0
Installs on skills.sh: 16,892 (Nov 2, 2026)
GitHub stars: 73,211 (Nov 2, 2026), whole repo
Files: 1 file including SKILL.md
Last commit: Oct 30, 2026 (repo)
Product: Strix, an open-source AI pentesting tool (Apache 2.0)
Option A needs: Docker on the runner, an LLM model id and API key as repository secrets
Exit codes: 0 pass, 2 vulnerabilities found, 1 setup error
Output: SARIF 2.1.0 for GitHub code scanning

What it includes

GitHub Actions workflow that runs strix in quick mode on the PR diff
Checks out with fetch-depth 0, installs Strix, runs a diff-scoped scan with a budget cap, then fails unless the run status is completed.
Runs code · Option A: self-hosted CLI
Stops a budget-capped run from passing silently
Reads run.json from strix_runs and fails the job if the status is not completed.
Runs code · Fail-unless-completed step
Optional upload to GitHub code scanning
Uses github/codeql-action/upload-sarif on the strix_runs folder.
Network · SARIF upload
Same steps for GitLab and others
Resolves the base branch, fails loudly if it cannot, and runs headless. Suggests standard scans nightly and deep scans for release candidates.
Runs code · Other CI systems
PR-review app, CLI-triggered review, or source upload
The app needs no workflow file. The CLI route uses strix cloud pr-reviews start with an API token. The source upload route uses a dry run first, then an approved upload.
Write · Option B: managed platform

From the official docs

Permissions

Medium risk

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

The pipeline pipes an installer into bash
The workflow runs curl -sSL https://strix.ai/install | bash on your runner. Pin or review the installer if your pipeline handles production secrets. github.com
Your code goes to an LLM provider or to Strix
Option A sends the code the scan analyses to the model provider you configure through STRIX_LLM and LLM_API_KEY. Option B reviews or uploads code on app.strix.ai; the source upload route has a dry run and a digest approval step first. Pick the option that fits your data rules. github.com
Every pull request run spends LLM credits
The Option A workflow runs a scan on each pull request with your provider key and `--max-budget 10`. A run that hits the budget can still exit 0, which is why SKILL.md adds a completion check. github.com
A passing scan is not proof of no bugs
The skill warns that a run that hits its budget can still exit 0. Keep the completion check and size the budget with room. github.com

Technical details

Talks to	strix.ai installer. your LLM provider (Option A). app.strix.ai (Option B). GitHub code scanning (SARIF upload)
Scopes	STRIX_LLM and LLM_API_KEY secrets (Option A) STRIX_API_TOKEN with pr_reviews:write (Option B CLI route)

Questions

What does this skill do?

It makes your agent add Strix security scanning to your CI pipeline. Each pull request gets a diff-scoped AI pentest, with findings as PR comments and SARIF for code scanning, and the job can fail on validated vulnerabilities.

How do I install it?

Use the install box above. It is one folder. The skill writes pipeline files: Option A installs Strix inside your CI job, and Option B's CLI route also uses the strix binary, including strix cloud repos list once to read the repository's provider and installation id. For the full Strix set the README suggests npx skills add usestrix/strix, which installs nine skills.

Does it need an API key?

Option A needs a model id and a provider API key stored as repository secrets named STRIX_LLM and LLM_API_KEY. Option B, the managed platform, needs no LLM key, but the CLI route uses a Strix API token.

Is it free?

The self-hosted CLI route is described as free with your own LLM key, so each pull request scan bills your model provider. The workflow it writes passes --max-budget 10; SKILL.md does not state the unit, so check Strix's docs before relying on it as a dollar cap. The managed platform is separate: SKILL.md says non-Enterprise cloud scans consume credits.

Is it safe?

The workflow it writes downloads and runs an installer with curl piped to bash in your CI runner, and sends code to your LLM provider or to Strix. Read the generated workflow, and keep the secrets in your CI secret store.

Which option should I pick?

The skill recommends the managed platform for most teams because there is nothing to maintain. It recommends the self-hosted CLI when scans must stay entirely inside your infrastructure.

How is it different from managed-pentesting-with-strix?

This skill sets up CI gating. managed-pentesting-with-strix covers the full strix cloud command line: scans, triage, reports, schedules and billing.

Information

updated Nov 2, 2026

Repository: usestrix/strix
Folder: skills/ci-security-scanning-with-strix

usestrix/strix skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 2

ci-security-scanning-with-strix SKILL.md · github.com
usestrix/strix README · github.com