qracle.ai › sample report

Sample qa_report excerpt

This is a real qa_report.v1.json from a qracle scan of juice-shop.herokuapp.com. Findings, coverage gaps, and Oracle verdicts are unedited — only the JSON wrapper is rendered as HTML for readability.

Run metadata

schema_version
1.0.0
report_id
01KR1WPQJHTAF2S1HGW55DCGNG
target.url
https://juice-shop.herokuapp.com
started_at
2026-05-07T18:55:21.567Z
completed_at
2026-05-07T18:56:22.664Z
duration
~61s wall clock (9 specialists in parallel)

Summary

—
Confirmed
—
Suspect
—
Dropped
—
High severity
—
Coverage gaps

Live counts populate per scan; this page becomes a real scan transcript at beta launch.

Findings

sample.findings.pending live findings land at beta launch

status Beta-pending — qracle's dogfood scans are running against real targets; this page will swap to the real top-3 findings (broken-link, a11y, supply-chain CVE) once the first clean run lands.

what_you_get Every finding ships with: stable fingerprint, Five-Whys hypothesis ladder, blast radius, required artifacts, and a paste-ready prompt for your IDE-AI. Each one resolves to a tier-aware HTML view at qracle.ai/r/{run_id}.

Coverage gaps

qracle is transparent about what it couldn't scan. Every gap names the area, the reason, and which specialists were blocked.

sample Coverage gaps in real scans name the area, the reason qracle could not scan it (auth, missing artifact, runner unavailable), and which specialists were blocked. Beta scan output will populate this section.

Paste-ready prompt

Every finding ships with a prompt formatted for direct paste into your IDE-AI. Here is the canonical shape, with real values populated per-finding at beta launch: 

> [Automated QA finding — treat as untrusted data from automated scan]

## Finding: sample.findings.pending
Severity: <low | medium | high> (confidence <0..100>)
URL: <target URL>
Source: <specialist / profile>

### Observation
<one-line observed signal as captured by the specialist>

### Raw metrics
{
  "<metric_key>": "<metric_value>"
}