Index of every page on this site: llms.txt
All skills
avivgerz-hue avatar

/ghost-qa

@19cc3ab
Run it yourself

Verify frontend work in a real headless browser before reporting a task complete. The autonomous Staff QA Engineer protocol — doctor readiness, multi-viewport 8-Trap baseline audit (Desktop/Tablet/Mobile, light + dark), DOM-to-source root-cause mapping, surgical remediation, and closed-loop delta verification with ghost_qa_compare until QA Health Score is 95–100 (Grade A+). Use after building or changing any UI, route, form, API integration, or client-side logic.

Use this Skill: https://skilld.dev/gh/avivgerz-hue/ghost-qa/ghost-qa

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SECURITY.md

≈495 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Security Policy

Supported versions

Version Supported
1.2.x ✅
1.1.x ✅ (security fixes only)
< 1.1 ❌

What ghost-qa is (and is not)

ghost-qa is a local development tool. It launches a headless browser against an app you are running and interacts with it like a cautious user would:

  • It only follows same-origin links from the target URL.
  • It refuses to click destructive controls (delete, checkout, pay, logout, publish, sign-in providers…) and never submits payments or orders.
  • Form fields are filled only with inert data (ghost-qa@example.com, ghost-qa test).
  • The MCP server speaks JSON-RPC over stdio only — it opens no network ports. Only connect it to MCP hosts you trust: anyone who can call ghost_qa_audit_url can make your machine browse URLs, so treat it like any other local agent tool.
  • --server-cmd executes the shell command you supply, in your shell. Never paste untrusted commands into it.

Do not point ghost-qa at production systems or websites you do not own or have explicit permission to test. Interaction testing clicks real buttons; on a system you don't control that may create data or trigger side effects.

Reporting a vulnerability

Please report security issues privately via GitHub's "Report a vulnerability" button (Security → Advisories) on this repository, rather than opening a public issue. Include:

  • Affected version (npx github:avivgerz-hue/ghost-qa --version)
  • A minimal reproduction (target fixture + command)
  • The impact you believe it has

We aim to acknowledge reports within 5 business days. Fixes ship on the latest minor version.

Scope notes

  • The accessibility trap checks common machine-checkable failures; it is not a WCAG conformance audit.
  • Performance thresholds are heuristics tuned for local development hardware; borderline findings on loaded CI machines should be re-run before being treated as defects.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

1 warning13d3 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    Ghost-qa is a headless browser verification tool for coding agents that crawls local applications to trap errors such as React hydration mismatches, network failures, and accessibility defects. The tool includes built-in safety mechanisms to prevent destructive interactions (e.g., clicking 'delete' buttons). A low-severity risk exists due to the Indirect Prompt Injection surface, where an agent could ingest malicious instructions embedded in a page being audited. It also supports execution of user-supplied shell commands to manage local development servers.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: MEDIUM · 1 issue

Signed by skilld at 19cc3ab. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 4 days ago.

Activeupdated 2 weeks ago

README badge

README badge for avivgerz-hue/ghost-qa