Output & Export

After a test, results are saved locally and can be viewed or exported in multiple formats.

Results Location

Results are saved to .humanbound/results/exp-{timestamp}/ in your working directory:

.humanbound/results/exp-20260419-135646/
    meta.json       # experiment metadata + posture + insights
    logs.jsonl      # conversation logs (one per line)

Viewing Results

Posture Score

hb posture

Shows the overall security posture score (0-100, grade A-F), defense rate, and breached categories.

Conversation Logs

# Table view (terminal)
hb logs

# Filter by verdict
hb logs --verdict fail

# Pagination
hb logs --page 2 --size 20

Export Logs

# JSON
hb logs --format json -o logs.json

# Interactive HTML
hb logs -f html -o logs.html

Reports

# HTML report (branded, self-contained)
hb report -o report.html

# JSON export
hb report --json -o results.json

The HTML report includes:

Cover page with test metadata
Posture score with donut chart
Stats cards (total/pass/fail/defense rate)
Severity distribution bar
Findings table with severity badges
Category coverage (breached vs defended)
Failed conversation appendix (up to 20)
Methodology description
Technology disclaimer

Guardrails Export

Export firewall rules from test findings:

# JSON format
hb guardrails -o rules.json

# YAML format
hb guardrails --format yaml -o rules.yaml

# OpenAI moderation format
hb guardrails --vendor openai -o openai_rules.json

Use with humanbound-firewall:

from hb_firewall import Firewall
fw = Firewall.from_config("agent.yaml", rules_path="rules.yaml")

Firewall Training

Train a Tier 2 classifier from test results:

# From local test data
hb firewall train

# From external results (vendor-agnostic)
hb firewall train --import pyrit_results.json
hb firewall train --import results.json:promptfoo

Log Schema

Each log entry contains 10 public fields:

Field	Type	Description
`thread_id`	string	Session identifier
`conversation`	array	Turns: `[{"u": "...", "a": "..."}]`
`result`	string	`pass`, `fail`, or `error`
`gen_category`	string	OWASP category tested
`fail_category`	string	Evaluation that triggered
`explanation`	string	Judge's verdict explanation
`severity`	float	0-100
`confidence`	float	0-100
`exec_t`	float	Avg execution time per turn (seconds)
`meta`	dict	Telemetry data (tools called, resource usage)

Experiment Metadata

meta.json contains the experiment summary:

{
  "id": "exp-20260419-135646",
  "name": "cli-owasp_agentic-20260419",
  "status": "Finished",
  "test_category": "humanbound/adversarial/owasp_agentic",
  "testing_level": "unit",
  "lang": "english",
  "results": {
    "stats": {"pass": 79, "fail": 18, "total": 97},
    "insights": [...],
    "posture": {"posture": 63.82, "grade": "C", ...},
    "exec_t": {"avg_t": 1.7, ...}
  },
  "created_at": "2026-04-19T13:56:46",
  "completed_at": "2026-04-19T14:15:22"
}