security-audit

A security review process for one MCP Runtime change. It examines the changed code and selects relevant local or continuous-integration checks.

In plain words
What is it for?
Reviewing authentication, permissions, secrets, TLS, Kubernetes settings, gateway rules, audit logging, dependencies, and GitHub Actions changes.
Why use it?
It helps catch security risks in a focused change without performing a full platform audit.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/agent-hellboy/mcp-runtime/security-audit
Any agent
npx skills add Agent-Hellboy/mcp-runtime --skill security-audit
Clone the repo
git clone --depth 1 https://github.com/Agent-Hellboy/mcp-runtime

Made for: Claude Code, Codex.

Per session 121 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,355 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00121 $0.01355
Opus 5 $0.00060 $0.00678
Sonnet 5 $0.00024 $0.00271
Haiku 4.5 $0.00012 $0.00136

Measured 2d ago against content hash c35c36ba1297, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

security-audit scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.codex/skills/security-audit/SKILL.md · 102 lines

How it starts

The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Security Audit (change-scoped)

Overview

Use this skill to perform a focused, diff-driven security review for an MCP Runtime change. Combine manual threat review with the repository's existing scanners, and keep findings grounded in concrete file and line references.

If the user asks for a deep, platform-wide audit, switch to security-audit-platform. For SBOM, signed images, GitHub Actions, or dependency-only work, use supply-chain-audit. For RBAC, Pod Security Standards, NetworkPolicy, and manifest hygiene, use k8s-hardening-audit.

Workflow

  1. Define the security boundary.

    • Inspect git status --short, git diff --stat, and diffs for touched files.
    • Identify whether the change affects trust boundaries: user input, JSON/YAML parsing, HTTP handlers, auth cookies or headers, Kubernetes API calls, shell commands, image references, registry endpoints, ingress hosts, TLS material, audit events, or CI workflows.
    • Read nearby tests and existing implementations before judging a pattern unsafe.
  2. Review MCP Runtime security invariants for the changed surface.

    • Secrets must come from Kubernetes Secrets, environment references, local test fixtures, or documented dev-only setup; do not hardcode real credentials or log token values. PLATFORM_ADMIN_PASSWORD is bootstrap-only and must not remain in steady-state env.
    • Admin and platform APIs must enforce the intended role, namespace, grant, session, and trust checks before mutating resources or returning sensitive data. The check is the requireRole(roleAdmin, …) wrap in each split service routes.go; absence on a mutating handler is a finding.
    • Gateway policy denies by default, scopes decisions to the rendered grant/session policy, and emits audit events for allow and deny paths without leaking credentials.
    • Public ingress and TLS changes preserve the documented registry.<domain>, mcp.<domain>, and platform.<domain> split; do not expose Grafana or Prometheus on the public platform host without an auth-aware design.
    • Registry and image changes respect dev HTTP versus production TLS behavior, image-pull-secret handling, and the single supported registry/registry-cert owner for registry/registry-tls.
    • Kubernetes manifests keep RBAC narrow, avoid privileged pods unless explicitly justified, preserve restricted pod-security labels in repo-owned namespaces, and avoid broad secret access.
    • HTML, JavaScript, logs, and API responses escape untrusted content and avoid reflecting secrets, auth headers, session cookies, JWTs, OIDC tokens, or API keys.
    • GitHub Actions and dependency changes keep actions pinned to commit SHAs, minimize token permissions, and avoid exposing repository or deploy secrets to untrusted PR code.

Read the full file on GitHub · 102 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 102 lines · 121 tokens per session scan A c35c36ba1297

Subscribe to this mod's changes

security-audit is a skill published in the GitHub repository Agent-Hellboy/mcp-runtime (5 stars, last pushed 8d ago), licensed Apache-2.0. It adds 121 tokens to every session and 1,355 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

local-frontend-check

Smoke-test or verify UI behaviour on the local Jarvis Registry frontend running at http://localhost/gateway. Use for manual regression checks, bug-fix verification, and end-to-end confirmation of specific flows without running the automated test suite.

ascending-llc/jarvis-registry · 52 tokens

generate-server-card

Generate an MCP server card JSON (mcp-gateway-registry format) by analyzing server source code in a folder or GitHub URL. Studies code to detect server name, tools, transport, auth, and generates a registry-compatible config.

agentic-community/mcp-gateway-registry · 52 tokens

debug

Debug issues in the MCP Gateway Registry using first-principles thinking. Invoke when something is broken, timing out, returning errors, or behaving unexpectedly. Forces structured root-cause analysis before any code change is proposed.

agentic-community/mcp-gateway-registry · 45 tokens

playwright-e2e-testing

Playwright modern end-to-end testing framework with cross-browser automation, auto-wait, and built-in test runner.

activepieces/activepieces · 29 tokens

build-audit-logs

Build or review audit trails in TypeScript/JavaScript apps using evlog (pipelines, typed actions, denials, retention, compliance-style reviews). For application code, not for extending the evlog package.

activepieces/activepieces · 49 tokens

triage-image-cves

Scan an Activepieces Docker image with grype for OS/base-image (deb) and application (npm) CVEs of High/Critical severity. Lists the 3 most-recent published tags and lets the user pick which to scan, validates each finding is real and reachable, and proves candidate fixes in an isolated git worktree (rebuild image +…

activepieces/activepieces · 136 tokens