Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    devguard

    1

    Evidence-first workflow (verified-dev v1.1.0) and token-saving verification gate with auto-detected .devguard/checks, gitleaks/osv-scanner security gates, and 28k-token subagent economics.

    $19

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    devguard

    devguard

    Example session with this skill installed

    Fix the duplicate webhook charge bug in src/billing/webhook.ts and verify the fix using devguard.

    • Read your context and instructions
    • Compiled the devguard

    Goal & Playbook Routing

    Routing to

    Playbook B (Bug Fix): Reproduce the duplicate webhook event on the billing store surface, isolate the missing idempotency guard in src/billing/webhook.ts:48, implement the smallest fix, and verify with a mutation check.

    Root Cause & Fix

    SurfaceFindingEvidence
    handleStripeWebhookprocessedEventIds.has(evt.id) was checked after chargeCustomer() instead of before(read src/billing/webhook.ts:48-63)
    Guard AddedMoved idempotency check to the top of handleStripeWebhook and recorded evt.id atomically(read src/billing/webhook.ts:44-51)

    Verification Receipts

    • Reproducer before fix: agent-verify webhook-repro -- npm test tests/webhook.test.ts -> exit=1 (duplicate event charged twice; log: .verify/webhook-repro.log)
    • Mutation check (guard removed): agent-verify webhook-mutation -- npm test -> exit=1 (log: .verify/webhook-mutation.log)
    • Full test suite with fix: agent-verify webhook-pass -- npm test -> exit=0 (14/14 passed; log: .verify/webhook-pass.log)

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

    About this skill

    Why Developers Use devguard (v1.1.0)

    Coding agents frequently claim a bug is fixed or a feature works based only on a clean compile or truncated shell output—and burn 3x your token quota by spawning subagents on small single-file tasks.

    devguard v1.1.0 installs an evidence-first engineering workflow (verified-dev) and token-saving verification gate across Claude Code, Cursor, VS Code (Copilot), Google Antigravity, OpenCode, Windsurf, Codex CLI, and Gemini CLI (macOS, Linux & Windows):

    • Real-Surface Verification Gate (.verify/<name>.log): Captures full test/runtime output and exact failure line numbers (errors_at=L...) into .verify/<name>.log via agent-verify instead of trusting compressed terminal summaries.
    • Auto-Detected .devguard/checks + Security Gates (gitleaks & osv-scanner): devguard gate init auto-detects your stack (package.json, Cargo.toml, Package.swift, *.py, *.sh) and layers self-skipping commit secrets (gitleaks staged scan) and push deps (osv-scanner lockfile CVE scan) gates.
    • 4-Step Token Escalation Ladder (~28k tokens/turn Rule): Enforces 0-token bash/ast-grep -> codegraph/projectmem -> <1,000-line inline read -> >=30,000-token single-shot batch subagent, keeping single-file edits and shared-lock builds inside the Main Agent.
    • 6 Playbooks + 30-Case LLM Eval & Cross-Model Review: Routes every task to Investigation, Bug Fix, Feature (with Pre-Mortem & 4-line Throughput Checkpoint), Refactor, Performance, or Prompt / LLM Pipeline Change (devguard eval run|compare + on-demand devguard review).

    Requires shell and file-system access to capture .verify/<name>.log receipts and run .devguard/checks. Companion security tools (gitleaks, osv-scanner) self-skip with an install hint when absent.

    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

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 6 days ago

    Requires

    Git
    Node.js
    tsc
    macOS

    Permissions

    Terminal / Shell
    Read Files
    Listed10 days ago
    Updated6 days ago

    What's inside

    Frequently Asked Questions

    devguard follows the universal SKILL.md standard and works out of the box with Claude Code, Cursor, VS Code Copilot, Windsurf, Codex CLI, and Gemini CLI.

    It enforces the 28k-token subagent bootstrap rule: single-file edits under 1,000 lines and known file:line fixes run directly in the main agent, avoiding ~3x subagent context overhead.

    New in Code Quality & Review