More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    ci cd Pipeline Validator

    3

    A junior dev pushes a workflow file with `uses: some-org/deploy-action@main` and `secrets.GITHUB_TOKEN` exposed.

    Free

    29 usesSecurity scanned
    ci-cd-pipeline-validator

    ci-cd-pipeline-validator

    Example session with this skill installed

    "Audit .github/workflows/deploy.yml for security and best practices"

    • Read your context and instructions
    • Compiled the ci-cd-pipeline-validator

    CI/CD PIPELINE AUDIT REPORT
    ═══════════════════════════

    Platform: GitHub Actions
    File: .github/workflows/deploy.yml

    Category Check Status Detail

    ─ ───────────── ────────────────────── ────── ─────────────────────────────────
    1 Security No hardcoded secrets PASS -
    2 Security Action versions pinned FAIL some-org/deploy@main → pin to SHA
    3 Security Script injection safe FAIL ${{ CI_ACCESS_TOKEN }} in run:
    4 Performance Cache configured WARN No dependency caching
    5 Best Practice Concurrency group WARN Parallel deploys can race
    6 Best Practice Timeout set WARN No timeout-minutes

    Top 3 fixes:

    1. [FAIL] Replace ${{ CI_ACCESS_TOKEN }} with env variable reference
    2. [FAIL] Pin deploy-action to commit SHA
    3. [WARN] Add concurrency group with cancel-in-progress

    Connects securely to your tools. The creator never sees your data.

    About this skill

    The Problem

    A junior dev pushes a workflow file with uses: some-org/deploy-action@main and secrets.GITHUB_TOKEN exposed. The PR merges. Now any fork can read your deploy secrets. You don't find out until a crypto-miner spins up on your runner. Or worse: the pipeline works for three months, then ubuntu-latest migrates from 22.04 to 24.04, and every build breaks on a Wednesday morning with no one knowing why.

    What You Get

    • Auto-detect the CI platform from file content — GitHub Actions, GitLab CI, CircleCI, or Jenkins — and apply platform-specific validation rules without manual selection
    • Scan for critical security anti-patterns — unpinned action references (@main/@master), script injection via ${{ github.event.* }}, pull_request_target misuse, hardcoded secrets, and overly permissive permissions: write-all
    • Validate YAML correctness before content — catch silent gotchas like unquoted version: 3.10 parsing as float 3.1, on: push vs on: [push] boolean parsing, and block scalar | vs > behavior differences
    • Optimize pipeline performance with dependency caching patterns (actions/cache@v4 with hashFiles keys), matrix build parallelism, and concurrency groups that prevent deploy pile-ups
    • Produce a prioritized audit report with PASS/WARN/FAIL status per check, a fix list ordered by exploitability, and before/after code snippets for every finding
    • Verify fixes don't break semantics — re-run the checklist post-fix, confirm no permissions were over-tightened, and validate the YAML still parses cleanly

    Why This Beats Prompting It Yourself

    Telling an agent "check my workflow file" produces surface-level feedback: "add caching" or "pin your actions." This skill runs a structured 5-category checklist (Syntax → Security → Performance → Best Practices → Platform-Specific) with exact pattern matching. It catches the compound vulnerabilities — like pull_request_target + checkout of PR head + secrets available — that a casual review misses entirely.

    Example Session

    You: "Audit .github/workflows/deploy.yml for security and best practices"

    Agent:

    CI/CD PIPELINE AUDIT REPORT
    ═══════════════════════════
    
    Platform: GitHub Actions
    File: .github/workflows/deploy.yml
    
    #  Category       Check                   Status  Detail
    ─  ─────────────  ──────────────────────  ──────  ─────────────────────────────────
    1  Security       No hardcoded secrets    PASS    -
    2  Security       Action versions pinned  FAIL    some-org/deploy@main → pin to SHA
    3  Security       Secret exposure   FAIL    ${{ CI_ACCESS_TOKEN }} in run:
    4  Performance    Cache configured        WARN    No dependency caching
    5  Best Practice  Concurrency group       WARN    Parallel deploys can race
    6  Best Practice  Timeout set             WARN    No timeout-minutes
    
    Top 3 fixes:
    1. [FAIL] Replace ${{ CI_ACCESS_TOKEN }} with env variable reference
    2. [FAIL] Pin deploy-action to commit SHA
    3. [WARN] Add concurrency group with cancel-in-progress
    

    Use Cases

    • Pre-merge review of PRs that modify CI/CD pipeline configurations
    • Security audit of workflows before enabling pull_request_target triggers
    • Migration checklist when moving from ubuntu-20.04 to ubuntu-24.04 runner images
    • Initial pipeline setup for new projects to avoid the 10 most common pitfalls
    • Post-incident review after a pipeline vulnerability is discovered
    • Standardizing CI/CD conventions across a team of 10+ repositories

    Known Limitations

    The validator is static — it cannot detect runtime failures (flaky tests, network-dependent steps, secret rotation needs). YAML syntax checking catches structural errors but not semantic issues like incorrect environment variable names. Multi-file pipeline configs (reusable workflows, includes, orbs) require scanning sibling files, not just the file the user pointed at.

    Upgrade to Pro

    Free flags unpinned actions, secret leaks, and YAML gotchas in a single workflow file. CI/CD Pipeline Validator Pro ($9.99) auto-fixes common findings with review-ready patches, estimates wasted CI spend in dollars, tracks pipeline maturity across 5 dimensions, and adds pre-commit hooks plus Azure DevOps support — get the Pro version at https://www.agensi.io/skills/ci-cd-pipeline-validator-pro

    The validator is static — it cannot detect runtime failures (flaky tests, network-dependent steps, secret rotation needs). YAML syntax checking catches structural errors but not semantic issues like incorrect environment variable names. Multi-file pipeline configs (reusable workflows, includes, orbs) require scanning sibling files, not just the file the user pointed at.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    29 uses

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Requires

    Python
    Git

    Permissions

    Write Files
    Network Access

    Allowed Hosts

    api.example.com
    Listed5 months ago
    Updated18 days ago

    What's inside

    Frequently Asked Questions

    The validator supports major CI/CD platforms including GitHub Actions, GitLab CI, CircleCI, and Jenkins (Jenkinsfile), automatically detecting the syntax and schema specific to each platform.

    Unlike basic linters that only check for syntax errors, this skill performs deep semantic analysis to find security vulnerabilities like unpinned actions and performance issues like missing caching strategies.

    Simply provide the content of your configuration file or a link to your repository; the agent will scan the code and return a structured report with prioritized findings and 'Before & After' remediation blocks.

    Security is a priority; the skill focuses on auditing the logic and configuration of your pipeline files and does not require access to your actual production environment or live secret values.

    Yes, the skill provides specific code snippets and configuration changes that you can copy and paste directly into your YAML or Groovy files to resolve the detected issues instantly.

    New in DevOps & Cloud