Changelog

Full changelog maintained at CHANGELOG.md.

Unreleased

AI provider/model routing closure — runtime adapters for OpenAI, Anthropic/Claude, DeepSeek, OpenRouter, Google Gemini and xAI/Grok behind the generic AI interface, with independent provider/model selection and truthful provider errors
Generic HTTP access/status differential capability through the finding lifecycle, with status-only changes never promoted
Authenticated CLI missions — session propagation into discovery/probes/tools and truthful rejection of failed login redirects
Autonomous mission lifecycle closure — continuation past the first finding, zero-finding honesty, artifact reconciliation and a secret-free event stream

v7.0.0 (2026-08-12)

Major Features

Clean Architecture v7 core (src/hunterx) — domain, application, infrastructure, engines, agents, tools, plugins, knowledge, reporting, config, CLI and API layers
Autonomous mission orchestration — create, run, checkpoint, resume and finalize full-spectrum security-assessment missions
Adaptive mission planning — attack-path planning, replanning and explainable next-best-action selection
Toolchain intelligence layer — 92 registered security tools with machine-readable contracts, structured execution, parsing/normalization and dependency-aware chaining
Evidence-driven vulnerability validation — hypothesis testing, validation verdicts, and controlled, safe proof/PoC engineering with replay verification
Proof & PoC Validation Engine — proof contracts, minimal safe proofs, replay, reproducibility, evidence-driven impact and confidence
Target memory & campaign intelligence — snapshots, diffs, coverage and revalidation planning
Cloud & SaaS attack-surface intelligence — provider detection, resource/exposure/environment classification and topology for AWS, Azure, GCP, OCI, Cloudflare and more
Knowledge graph & correlation — cross-tool evidence chains and attack-path analysis
Professional reporting — findings, evidence bundles, remediation plans and multi-format exports (Markdown, HTML, JSON, SARIF, PDF, package)
TIDB persistence — SQL storage with 21 linear Alembic migrations, events, audit and versioning
HUNTERX_* environment-variable configuration overrides
Mission persistence across CLI invocations and process restarts (restore path)
Safe XML parsing via defusedxml (XXE / entity-expansion hardened)
install.sh v7 installer (idempotent, database initialization, verified installation)

Fixed

Base installation no longer crashes without optional extras (SQLAlchemy now a base dependency)
mission create → mission start workflow works across separate CLI invocations
Harden hunterx base install and Docker image runtime
Ruff, dead-code (vulture) and bandit security gates green

Engineering Validation

3479 tests passed, 8 skipped, 2 deselected, 0 failed
Ruff, mypy, bandit, vulture, docs and package gates green

v6.0.0 (2026-07-26)

Major Features

Security Skills Framework — 41 built-in skills with MITRE ATT&CK, OWASP, CWE, CAPEC mapping
Reasoning Engine — 18 goal types with planner, prompt builder, validator, and multi-call consensus
Multi-Agent Platform — 10 specialized agents with event/message buses, DAG workflows, checkpoint/resume
AI Provider Abstraction Layer — OpenAI and Ollama providers with session management, caching, metrics, middleware, retry, circuit breaker
Knowledge Graph — Entity-relationship store for findings, targets, attack paths, and contextual data
Threat Modeling & Attack Chains — STRIDE/LINDDUN categorization, trust boundary mapping, chain analysis
Payload Intelligence — SQLite + FTS5 indexing, 5-level execution policy, 10 mutation families, provenance tracking, feedback loop
MITRE ATT&CK Mapping — All skills and findings mapped to Enterprise techniques

REST API (40+ endpoints)

Scan lifecycle: submit, poll, cancel, pause, resume
AI provider management: providers, chat, streaming, sessions, metrics
Agent management: list, details, start, stop
Skills: list, search, execute, install, uninstall, categories, MITRE mapping
Payload: search, stats, mutate, techniques, feedback
Workflows: list, create, execute, status
Reasoning: create goals, get results
System: health, config, version

CLI

Commands: scan, module, report, doctor, config, update, api
Subcommands: skills, agents, payload, workflow, reasoning, ai
Profile support: internal, bounty, gov
Preset support: quick, full, stealth
Authentication: basic, bearer, cookie, form, JWT
AI integration flags with multi-provider support

Reporting

JSON, Markdown, SARIF 2.1, HTML
Visual attack graphs (HTML + Graphviz)
Purple team detection rules
ZIP evidence packages

Testing & Quality

623 passing tests
0 Ruff errors
CI matrix for Python 3.11/3.12/3.13

v4.0.1 (2026-07-22)

License Change

License changed from Proprietary to Apache 2.0 — HunterX is now fully open source
All source files updated to SPDX-License-Identifier: Apache-2.0
CONTRIBUTING.md updated with DCO (Developer Certificate of Origin) requirement

Documentation

Comprehensive README with badges, CI status, and architecture
GitHub Pages site with SEO metadata and structured data
Quickstart guide, API reference, Docker guide, FAQ, tutorials

Infrastructure

GitHub Actions CI pipeline with Ruff linting and Pytest
Docker multi-stage build producing 271MB image
Docker Hub automated builds

v4.0.0 (2026-07-20)

Initial Release

4-stage reasoning pipeline (Passive Intel, Probe, Confirm, Verify)
200+ vulnerability signatures across web, API, and cloud categories
REST API server with async job queue
Authentication support (Basic, Bearer, Cookie, Form Login)
Plugin system with detectors, reporters, and hooks
Reporting: Markdown, JSON, SARIF 2.1, HTML, ZIP
WebSocket and GraphQL protocol testing
OOB detection for blind XXE, SSRF, and RCE
WAF fingerprinting with 50+ signatures
Operator profiles (internal, bounty, gov)
Stealth modes with configurable timing
76 passing tests