Agent Skills — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Agent Skills (MCP Server) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Teach Claude Code, Cursor, and other agents how to use Zephex MCP and the zephex terminal CLI on the user's own repository.
npx skills add zephexMCP/agent-skills
# or
npx skills add https://github.com/zephexmcp/agent-skills --skill zephex| Need | How | |
|---|---|---|
| MCP in editor | npx zephex setup (or copy from configs/) | |
| API key | zephex.dev/dashboard/api-keys | |
| Terminal CLI (optional) | `curl -fsSL https://zephex.dev/cli/install.sh \ | bash` |
https://zephex.dev/cli/install.sh and https://zephex.dev/install.sh are the same installer. Windows: irm https://zephex.dev/install.ps1 | iex.
| File | Role |
|---|---|
skills/zephex/SKILL.md | Full routing: 10 tools + CLI commands |
AGENTS.md | Short session gate |
CLAUDE.md | Claude Code preferences |
configs/* | Optional manual MCP templates |
Deep public catalog (examples + CLI map): [zephex-tools](https://github.com/zephexMCP/zephex-tools) · On-site: docs/tools
| Tool | Purpose |
|---|---|
get_project_context | Framework, scripts, env, monorepo (topic slices) |
find_code | BM25 + ripgrep search |
read_code | AST symbol / file / outline |
explain_architecture | Wiring map + Mermaid-capable deep mode |
check_package | Registry safety, CVEs, upgrades (12 ecosystems) |
check_test | Test Pulse — run suite + failure health |
audit_headers | Live HTTPS URL audit |
project_memory | Remember / recall project facts |
keep_thinking | Multi-step reasoning + loop detection |
Zephex_dev_info | Expert implementation playbooks |
curl -fsSL https://zephex.dev/cli/install.sh | bash
cd your-project && zephex login
zephex deep --json # orientation packet for agents
zephex structure --agent # free folder map
zephex architecture --focus auth
zephex overview
zephex find "auth middleware"
zephex test # run suite
zephex check test failures # after test
zephex check test fix-prompt --copy
zephex safe lodash
zephex learn # free catalognode scripts/eval-description.mjs (20/20)Ground truth: mcpHub available-tools.ts + proxy tool schemas.
There is no “publish” button. skills.sh indexes via install telemetry + cached snapshots. GitHub main is the source of truth; the catalog Overview/Snyk can lag for weeks (#919, #1273).
| Check | Expected (current product) |
|---|---|
| raw SKILL.md | Call order starts with get_project_context → find_code; no scope_task / inspect_url |
| download API | Should match GitHub after reindex (stale if it still lists references/tools.md) |
| Snyk page | May still show May 2026 W011 until re-audit |
Install always clones GitHub, not the skills.sh HTML cache:
npx skills add zephexMCP/agent-skills --skill zephex -y
npx skills update zephex -yNudge reindex: reinstall via CLI, then if download API stays stale, open an issue on vercel-labs/skills.
See LICENSE.
| Repo | Role |
|---|---|
| zephex-MCPs | Product overview |
| zephex-tools | Ten tools deep catalog |
| zephex-cli | Mode 2 CLI |
| zephex-web-terminal | Browser terminal |
| llms.txt | Agent product truth |
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.