Quick Start — Oathe Docs

Submit your first behavioral security audit in under 3 minutes. No API key, no signup.

1. Pick a skill to audit

Any public GitHub repository that contains an MCP server, plugin, or agent skill. Example:

https://github.com/anthropics/model-context-protocol

Option A — Website

Go to oathe.ai, paste the URL into the audit bar, and press Scan.

Option B — cURL

curl -X POST https://audit-engine.oathe.ai/api/submit \
  -H "Content-Type: application/json" \
  -d '{"skill_url": "https://github.com/owner/repo"}'

You’ll get back an audit_id:

{
  "audit_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "queue_position": 1
}

Option C — MCP Server

npx oathe-mcp

Then call submit_audit with the skill URL. See the MCP Server guide for details.

Audits take 30–90 seconds. The skill is installed in an isolated environment, its behavior is monitored, and threat patterns are applied.

Poll for status:

curl https://audit-engine.oathe.ai/api/audit/{audit_id}

Or watch the progress bar on the website — it uses WebSocket for real-time updates.

When status is "complete", the response includes a full report:

Field	What it tells you
`recommendation`	What to do: INSTALL, INSTALL_WITH_CAUTION, REVIEW_BEFORE_INSTALL, or DO_NOT_INSTALL
`trust_score`	0–100 composite score
`verdict`	SAFE, CAUTION, DANGEROUS, or MALICIOUS
`findings`	Specific security issues detected
`summary`	One-sentence assessment

The recommendation field is the single most important signal for decision-making.