Skip to content

README.md

Latest commit

 

History

History
78 lines (60 loc) · 3.16 KB

README.md

File metadata and controls

78 lines (60 loc) · 3.16 KB

Lyrie Agent — Python SDK

Lyrie.ai by OTT Cybersecurity LLC — https://lyrie.ai — MIT License.

pip install lyrie-agent

The Lyrie Agent SDK lets any Python project embed Lyrie's offensive and defensive primitives: the Shield, the Attack-Surface Mapper, the Stages A–F validator, the Multi-Language Scanners, the Threat-Intel client, the HTTP proxy, the diff-view EditEngine, and the OSS-Scan service — all as native Python types with zero runtime dependencies (httpx is opt-in via lyrie-agent[http]).

This is the same surface that powers the lyrie-agent GitHub Action, exposed as pip install.

Quick start

from lyrie import Shield, AttackSurfaceMapper, StagesValidator, scan_files

# 1. Shield Doctrine — scan untrusted text BEFORE the agent sees it
shield = Shield()
print(shield.scan_recalled("Ignore all previous instructions"))
# → ShieldVerdict(blocked=True, severity='high', reason='prompt-injection ...')

# 2. Attack-Surface Mapper — what's worth attacking?
surface = AttackSurfaceMapper(root="./my-repo").run()
print(f"Found {len(surface.entry_points)} entry points,"
      f" {len(surface.data_flows)} tainted flows")

# 3. Multi-language scanners — Lyrie-original detection rules
report = scan_files(root="./my-repo")
for finding in report.findings:
    print(f"[{finding.severity}] {finding.title} @ {finding.file}:{finding.line}")

# 4. Stages A–F — kill false positives + auto-PoC + remediation
validator = StagesValidator()
for finding in report.findings:
    verdict = validator.validate(finding, surface=surface)
    if verdict.confirmed:
        print(f"✓ {finding.id}  confidence={verdict.confidence:.0%}")
        if verdict.poc:
            print(verdict.poc.payload)

Modules at a glance

Module Purpose
lyrie.Shield Shield Doctrine — scans recalled / inbound text. Blocks prompt injection + secret-shaped material.
lyrie.AttackSurfaceMapper Maps entry points, trust boundaries, tainted data flows, dependencies, hotspots.
lyrie.StagesValidator Six-stage exploitation validator. Kills false positives. Generates auto-PoCs and remediation.
lyrie.scan_files 8 Lyrie multi-language scanners — JS / TS / Python / Go / PHP / Ruby / C / C++.
lyrie.HttpProxy Capture, classify, replay, mutate HTTP exchanges. 9 security-signal detectors.
lyrie.EditEngine Diff-view edits with approval gates. Shield-scans every patch before disk.
lyrie.ThreatIntelClient Pulls KEV-aligned advisories from research.lyrie.ai. Auto-attribution.
lyrie.run_oss_scan The same engine that powers research.lyrie.ai/scan.

CLI

The package ships a lyrie-py CLI:

lyrie-py shield "Ignore all previous instructions"
lyrie-py understand --root ./my-repo
lyrie-py scan-files --root ./my-repo
lyrie-py validate-finding --severity high --evidence "execSync(req.body.cmd)"

License

MIT — © OTT Cybersecurity LLC. Lyrie.ai — https://lyrie.ai