AXguard

Open-source AI security tool to scan and fix vulnerabilities in your vibe-coded apps before you ship.

AXguard by AwareXone

Built by Shuvonsec

AXguard is built by Shuvonsec — Ethical hacker and security researcher. He ranked #1 worldwide on the TryHackMe monthly leaderboard in 2025. He works on AI security and cybersecurity agents, and builds open-source tools to make AI-built applications safer.

CLI + AI agent plugin for Claude Code, Cursor, OpenCode, Codex, and shared Agent Skills.

Pre-ship security gate — not a full pentest platform. Scan source, triage noise, fix what matters, and gate releases before you publish.

Website · GitHub · X · Dev Docs · Commands

Contact: [email protected] · [email protected] · [email protected]

---

What is AXguard?

AXguard is a pre-ship security gate.

Before you publish an app — especially one built with AI — AXguard checks your code for common security problems, helps you fix them, and writes clear reports.

It ships as:

• A standalone CLI (axguard)
• An AI agent plugin (skills + slash commands for Claude Code, Cursor, OpenCode, Codex, and shared Agent Skills)
• An optional local-first Security Intelligence API (axguard api start → http://127.0.0.1:8787) — no AwareXone account or hosted LLM; use no-llm (default), Ollama/local, or BYOK (pip install -e '.[api]'). See docs/api/overview.md.

AI builds it → AXguard checks it → You fix it → You ship it

Source scanning is current. Artifact/bytecode scanners (JS bundles, WASM, etc.) are planned — see CONTRIBUTING.md.

---

Why AXguard?

AI tools can build an app in minutes. They can also ship security bugs in minutes.

AXguard sits between:

"the AI built it"
        ↓
     AXguard
        ↓
"we shipped it"

Built for developers, founders, security engineers, and teams that want a simple security checkpoint — without a heavy enterprise setup.

---

Workflow

threat-model → audit → triage → fix → re-scan → report → CI

Step	What happens
threat-model	Map the app, trust boundaries, and likely risks before a deep pass.
audit	Run the full pre-ship scan and write HTML / MD / JSON plus diagnostics.
triage	Drop false positives; keep confirmed and likely issues.
fix	Patch confirmed bugs with remediation guidance.
re-scan	Re-run scan/audit to verify the fix held.
report	Produce a shareable security report for PRs and stakeholders.
CI	Fail the pipeline on high/critical so regressions do not ship.

---

Quick Start

1. Install the plugin

git clone https://github.com/Awarexone/AXguard.git
cd AXguard

chmod +x install.sh uninstall.sh
./install.sh

For Cursor:

./install.sh --agent cursor

For all supported agents:

./install.sh --agent all

Supported install targets (./install.sh --agent …): claude, cursor, opencode, codex, agents (shared Agent Skills), and all.

2. Install the CLI

python3 -m venv .venv
source .venv/bin/activate

pip install -e .

axguard help
axguard audit .

Open the report:

open .findings/axguard/axguard-report.html

Package	axguard
Version	0.2.0
Python	3.10+
License	MIT

---

Start Using AXguard

Start with the workflow you need, not the full list.

What are you doing?	Start here
About to ship	/axguard-audit
Quick check while coding	/axguard-scan
New or unknown codebase	/axguard-threat-model
Map attack surface / app model	/axguard-surface
Dataflow / taint paths	/axguard-flow
Hunter → Judge verification	/axguard-verify
False-positive adversary	/axguard-adversary
Evidence & confidence	/axguard-evidence
Attack graph / vuln chaining	/axguard-paths
AI-generated / agent app	/axguard-agent
Looking for leaked keys	/axguard-secrets
Auth / IDOR issues	/axguard-auth
Injection / RCE	/axguard-inject
SQL injection	/axguard-sql
SSTI	/axguard-ssti
Path traversal / LFI	/axguard-path
SSRF	/axguard-ssrf
XSS	/axguard-xss
Cloud / CORS	/axguard-cloud
Crypto misuse	/axguard-crypto
Supply chain	/axguard-supply
GraphQL	/axguard-graphql
File upload	/axguard-upload
Debug / error leaks	/axguard-debug
Too many findings	/axguard-triage
Fix confirmed bugs	/axguard-fix
Need a report	/axguard-report
Add a CI gate	/axguard-ci
Training-data pipeline	/axguard-data
Contribute (local, opt-in)	/axguard-contribute
Privacy prefs (local)	/axguard-privacy
Security Twin	axguard twin build . → docs/twin
Security Memory	axguard memory record . → docs/memory
Investigation Agent	axguard investigate . → docs/investigation
Predictive security risk	axguard predict . → docs/predictive
Local Security Intelligence API	axguard api start → docs/api
GitHub PR bot (self-host)	axguard github setup → docs/github
Full security-lead pass	skill axguard-cso
Short pre-ship checklist	skill axguard-preship

See the command cheat sheet.

---

Commands & Specialists

Each command has a clear job. Slash commands install via ./install.sh; most also have a CLI equivalent.

Command	Specialist	Key CLI	What it does
/axguard-audit	Pre-ship Lead	axguard audit .	Full audit + HTML / MD / JSON (+ diagnostics)
/axguard-scan	Scanner	axguard scan .	Fast check while you code
/axguard-threat-model	CSO	—	Map risks before a deep scan
/axguard-surface	Surface Mapper	axguard surface .	Application model (routes, sinks, stack)
/axguard-flow	Dataflow	axguard flow .	Dataflow / taint paths (diagnostic)
/axguard-verify	Verifier	axguard verify .	Hunter → Judge verification (diagnostic)
/axguard-adversary	FP Adversary	axguard adversary .	False-positive adversary (diagnostic)
/axguard-evidence	Evidence	axguard evidence .	Evidence & confidence (diagnostic)
/axguard-paths	Attack Graph	axguard paths .	Attack graph + vuln chaining (diagnostic)
/axguard-secrets	Secrets Hunter	—	Find keys and credentials
/axguard-auth	Access Control	—	Auth, IDOR, JWT, CSRF
/axguard-inject	Injection Hunter	—	Injection and RCE patterns
/axguard-sql	SQL Hunter	—	SQL / ORM injection sinks
/axguard-ssti	Template Hunter	—	Server-side template injection
/axguard-path	Path Hunter	—	Traversal / LFI / dynamic includes
/axguard-ssrf	Egress Hunter	—	Unsafe outbound requests
/axguard-xss	Client Security	—	Dangerous XSS sinks
/axguard-cloud	Cloud Reviewer	—	Cloud and CORS issues
/axguard-crypto	Crypto Reviewer	—	Weak hashing, hard-coded keys, TLS verify-off
/axguard-supply	Supply Chain	—	Install hooks, curl|sh, untrusted indexes
/axguard-graphql	GraphQL Reviewer	—	Introspection / CSRF footguns
/axguard-upload	Upload Hunter	—	Unsafe file upload patterns
/axguard-debug	Debug Hunter	—	DEBUG mode, stack traces, actuators
/axguard-agent	Agent Security	—	AI-agent and LLM risks
/axguard-triage	Triage Lead	—	Cut noise, keep real issues
/axguard-fix	Remediation	—	Fix and re-check
/axguard-report	Report Author	—	Clean security reports
/axguard-ci	Release Gate	axguard audit . --fail-on high	Fail CI on high / critical
/axguard-data	Data Pipeline	axguard data …	Training-data registry & pipeline
/axguard-contribute	Contributor	axguard contribute …	Local contribution suggest/prepare (no auto-push)
/axguard-privacy	Privacy	axguard privacy …	Local opt-in / export / delete prefs
skill axguard-cso	Chief Security Officer	—	End-to-end security pass
skill axguard-preship	Release Reviewer	—	Short pre-ship checklist

---

Which workflow?

Situation	Start with	Then
Shipping soon	/axguard-audit	/axguard-triage → /axguard-fix → re-scan
AI-built app	/axguard-agent	/axguard-audit
Auth-heavy API	/axguard-auth	/axguard-audit
New codebase	/axguard-threat-model	/axguard-surface → /axguard-audit
Noisy results	/axguard-triage or /axguard-adversary	/axguard-fix
Need a shareable report	/axguard-report	Open the HTML
Want CI protection	/axguard-ci	Add it to your pipeline

---

CLI

No AI agent required.

axguard help
axguard version
axguard about
axguard engage disable|enable|dismiss

# Scan & audit
axguard scan .
axguard scan . --format json -o out.json
axguard audit .
axguard audit . --fail-on high
axguard audit . --fail-on high --out-dir .findings/axguard --no-banner
axguard audit . --open-summary

# Diagnostics (not vuln reports)
axguard surface .
axguard flow .
axguard verify .
axguard adversary .
axguard evidence .
axguard paths .              # alias: axguard attack-paths .
axguard paths . --current --critical --shortest
axguard paths . --predictive
axguard paths . --what-if remove_authz

# Security Twin (symbolic model; local)
axguard twin build .
axguard twin show .
axguard twin attack .
axguard twin blast-radius . --entity ENTITY
axguard twin controls .
axguard twin what-if .
axguard twin compare .
axguard twin regression .
axguard twin query .
axguard twin scenarios
axguard twin export-dataset .

# Security Memory (longitudinal; local)
axguard memory record .
axguard memory show
axguard memory history
axguard memory changes
axguard memory regressions
axguard memory findings
axguard memory controls
axguard memory paths
axguard memory unknowns
axguard memory query "what changed"

# Investigation Agent
axguard investigate .
axguard investigate . --fast
axguard investigate . --deep
axguard investigate . --finding FINDING_ID
axguard investigate --explain FINDING_ID

# Predictive security (risk signals — not confirmed findings)
axguard predict .
axguard predict --pr --base ./base-checkout
axguard predict --architecture
axguard predict --agent
axguard predict --mcp
axguard predict --what-if

# Training-data pipeline (no model training)
axguard data discover
axguard data inspect
axguard data report fixtures/data_pipeline

# Contribute / privacy (local prefs; no telemetry by default)
axguard privacy status
axguard privacy opt-in
axguard privacy opt-out
axguard privacy export
axguard contribute status
axguard contribute suggest
axguard contribute prepare

# Optional GitHub Security Bot adapter
axguard github setup .
axguard github validate .
axguard github test .
axguard github status .

# Local Security Intelligence API (optional extra: pip install -e '.[api]')
axguard api start
axguard api status
axguard api projects
axguard api keys list

open .findings/axguard/axguard-report.html

Common flags on scan/audit: --fail-on {critical,high,medium,low,none}, --rules DIR, --no-banner, --no-engage. Audit also accepts --out-dir and --open-summary. Scan also accepts --format {text,json,md,html} and -o / --output.

Docs: docs/twin · docs/memory · docs/investigation · docs/predictive · docs/api · docs/github · docs/contributors · docs/engagement.md

---

What it finds

Detections come from rules/*.json (deterministic pattern/heuristic rules).

Area	Examples (from shipped rules)
Secrets	AWS access keys, hard-coded API key assignments, PEM private keys, GitHub PATs, Slack tokens
Auth	Missing ownership heuristics (IDOR-style), JWT none, JWT decode-without-verify, CSRF disabled
Injection / RCE	eval / exec, pickle.loads, subprocess(shell=True), os.system / popen, JS child_process.exec, PHP unserialize, Java ObjectInputStream
SQL / NoSQL	String-built SQL (Python/JS/PHP), ORM raw SQL, Mongo $where / operator injection
SSTI	Jinja2 / Flask render_template_string, dynamic Pug compile
Path / LFI	User paths in open / send_file / PHP include
SSRF	Variable URLs in Python HTTP clients / JS fetch concatenation
XSS	innerHTML, document.write, React dangerouslySetInnerHTML
Upload	Original client filenames saved, Multer .any()
Crypto	Hard-coded keys/IVs, MD5/SHA1 near passwords, Math.random tokens, TLS verify off
Supply chain	Risky npm install scripts, pip extra indexes, curl|sh
GraphQL	Introspection left on, CSRF prevention off
Debug	Django/Flask debug, Express stack handlers, Spring Actuator hints
Cloud	AWS metadata URLs, wildcard CORS + credentials, public S3 ACL
AI agents	Executing model/agent output, unrestricted shell tools, prompt/tool-arg shell execution

---

Reports

Every full audit writes primary vulnerability reports:

.findings/axguard/
├── axguard-report.html   # easy to read
├── axguard-report.md     # for PRs and docs
└── axguard-report.json   # for CI and tools

The same audit also writes diagnostic artifacts (not the primary vuln report):

.findings/axguard/
├── application-model.{json,md}   # surface / app model
├── dataflow.{json,md}            # taint / dataflow paths
├── verification.{json,md}        # Hunter → Judge
├── adversary.{json,md}           # false-positive adversary
├── final-findings.json           # post-adversary finding statuses
├── evidence.{json,md}            # evidence & confidence
├── attack-paths.{json,md}        # attack graph / chaining
├── memory/                       # soft Security Memory snapshot (best-effort)
└── investigation/                # soft Investigation Agent pass (best-effort)

A full audit may also embed a twin summary in the main report; full security-twin.* files come from axguard twin … (default under .findings/axguard/twin/). Dedicated commands can write more under predictive/, data/, and contribute/.

---

CI / Release Gate

Code → AXguard → High/Critical?
                 ├── Yes → Fix → Re-scan
                 └── No  → Ship

Example for your app:

axguard audit . --fail-on high --no-banner

What this repository runs

.github/workflows/axguard.yml dogfoods runtime code only (engines/ and cli/). Fixtures, skills, commands, and rules intentionally contain vulnerable examples for regression tests and are excluded:

axguard audit engines --fail-on high --no-banner
axguard audit cli --fail-on high --no-banner

---

GitHub Security Bot

Optional GitHub App that reviews pull requests with Check Runs and one updatable summary comment. Runs AXGuard Core behind a thin webhook adapter (engines/github/). Local CLI scanning does not require it.

Add the bot to a repository

A. GitHub App (Check Runs + PR comment) — preferred when you can self-host:

1. Create a GitHub App with the permissions in docs/github/permissions.md.
2. Install it on Only select repositories (e.g. this repo).
3. Point the webhook at your adapter and export credentials (never commit them):

export AXGUARD_GITHUB_APP_ID=…
export AXGUARD_GITHUB_WEBHOOK_SECRET=…
export AXGUARD_GITHUB_PRIVATE_KEY_PATH=/path/to/app.pem

1. In the target repo:

axguard github setup .          # writes .axguard.yml (no secrets)
axguard github validate .
# run the adapter — see docs/github/self-hosting.md
axguard github status .

B. Actions-only (no webhook host) — what this repository uses in CI:

# .github/workflows/axguard.yml on pull_request / push to main:
axguard audit engines --fail-on high --no-banner
axguard audit cli --fail-on high --no-banner

That path runs Core in CI; it does not replace App Check Runs unless you also wire the App.

CLI

axguard github setup .
axguard github validate .
axguard github test .
axguard github status .

• Install & permissions: docs/github/install.md · docs/github/permissions.md
• Config (.axguard.yml): docs/github/config.md
• Self-host (preferred): docs/github/self-hosting.md
• Privacy / AI: docs/github/privacy.md · docs/github/ai-providers.md
• Architecture research: docs/research/github-security-bot.md
• Marketplace prep only (no approval claimed): docs/github/marketplace.md

---

Predictive Security

Predict security risk expansion from observable changes — not CVEs, not guaranteed future bugs. Separates Verified Issues from Predictive Risks.

axguard predict .
axguard predict --pr --base ./base-checkout
axguard predict --architecture
axguard predict --agent
axguard predict --mcp
axguard predict --what-if

• Guide: docs/predictive/README.md
• Research: docs/research/predictive-security.md

---

Built for vibe-coded apps

AI can generate an app quickly. Security review should still happen before you publish.

AI builds it
     ↓
AXguard checks it
     ↓
You fix it
     ↓
AXguard checks again
     ↓
You ship it

Use /axguard-agent when the app gives models access to shells, files, APIs, or tools.

---

AwareXone

AXguard is built by AwareXone.

We build open-source security tools for the AI era — for people who build and people who hunt.

BUILD                         HUNT
  │                             │
AXguard                 Agentic Bug Hunter
  │                             │
Secure what you create   Find bugs that are live

Same security DNA. Different job.

Agentic Bug Hunter

AI-powered bug bounty toolkit (4.8k+ stars).

Point it at a live target. It helps you recon, find vulnerabilities, validate findings, and write reports. Claude Code plugin + standalone bughunter CLI.

→ github.com/Awarexone/Agentic-Bug-Hunter

More from AwareXone

Tool	What it is
Agentic Bug Hunter	AI bug bounty toolkit
Public Skills Builder	Turn public security research into reusable skills
Web3 Bug Bounty AI Skills	Smart-contract and DeFi security skills

awarexone.com · GitHub · X @AwareXone

Beyond open-source tools, AwareXone also builds AI-driven defenses against scams, fraud, and social engineering, and provides human-risk security services for organizations.

General	[email protected]
Business / B2B	[email protected]
Founder	[email protected]

→ Get in touch

---

Project structure

AXguard/
├── cli/                 # axguard CLI entrypoints
├── engines/             # scanners, diagnostics, adapters
├── rules/               # detection rule packs (*.json)
├── commands/            # slash-command markdown
├── skills/              # orchestration + domain skills
├── fixtures/            # vuln / safe regression apps
├── tests/               # pytest suite
├── docs/                # architecture, API, twin, memory, …
├── references/          # framework / dataset provenance
├── scripts/             # validation helpers
├── data/                # data-pipeline seeds / fixtures
├── assets/              # README images
├── install.sh           # agent plugin installer
├── uninstall.sh
├── pyproject.toml
├── COMMANDS-QUICK-REF.md
├── DEV.md
├── CONTRIBUTING.md
└── LICENSE

---

Developer Docs

Doc	For
DEV.md	Setup and day-to-day development
COMMANDS-QUICK-REF.md	Slash + CLI cheat sheet
docs/architecture.md	How the scanner works
docs/attack-graph.md	Attack-graph diagnostics
docs/github/README.md	GitHub Security Bot (App adapter)
docs/predictive/README.md	Predictive security intelligence
docs/api/overview.md	Local Security Intelligence API
docs/twin/README.md	Security Twin
docs/memory/README.md	Security Memory
docs/investigation/README.md	Investigation Agent
docs/data/README.md	Training-data pipeline
docs/contributors/README.md	Contribute / privacy (local, opt-in)
docs/engagement.md	Engagement prefs (no telemetry)
docs/adding-rules.md	Adding detections
docs/plugin.md	Agent plugin setup
docs/SKILL-SCHEMA.md	Domain skill frontmatter + sections
docs/SECURITY-KNOWLEDGE-INVENTORY.md	Knowledge-layer inventory
skills/index.yaml	Skill registry (30 core + orchestration)
references/	Framework / repo / dataset provenance
CONTRIBUTING.md	Contributing to AXguard
CONTRIBUTORS.md	Contributors
skills/axguard-knowledge/	Vuln-class knowledge pack

---

Security knowledge layer

AXguard pairs a deterministic CLI scanner with a research-backed skill system so agents can reason, not just match regexes.

Frameworks (OWASP / CWE / …)
        ↓
references/ (provenance index)
        ↓
skills/security/* (30 core domain skills)
        ↓
skills/axguard-* (orchestration: audit → triage → fix → report)
        ↓
commands/ + rules/ + CLI

• Orchestration skills drive workflows (/axguard-audit, triage, remediate).
• Domain skills teach source→sink analysis, evidence gates, FP controls, and fixes per class (SSRF, SQLi, authZ, prompt injection, MCP, …).
• Provenance lives in references/ — cite official IDs only; no invented CWE/OWASP mappings; HF datasets are metadata/derived-knowledge only.
• Validate with: python scripts/validate_skills.py

---

Security checks

AXGuard checks your apps before ship. This repository also runs automated checks on itself:

Check	Workflow
CI tests + fixture self-scan	ci.yml
AXGuard Security Review (PR bot / Core)	axguard.yml
CodeQL (Python)	codeql.yml
Secret detection (Gitleaks)	gitleaks.yml
Dependency vulns (OSV-Scanner)	osv-scanner.yml
Actions audit (zizmor)	zizmor.yml
OpenSSF Scorecard	scorecard.yml
Dependency updates	Dependabot

Report vulnerabilities in AXGuard via SECURITY.md (GitHub Private Vulnerability Reporting preferred).

---

Contributing

Contributions are welcome.

Help with:

• Detection rules
• Scanners
• Test fixtures
• Agent skills
• Reports
• Documentation
• Bug fixes

pip install -e ".[dev]"
pytest -q
axguard audit fixtures/vuln_app --no-banner

See CONTRIBUTING.md. Artifact/bytecode scanners are on the roadmap there — not claimed as current features.

Local contribution packaging is opt-in and never auto-pushes:

axguard privacy opt-in
axguard contribute suggest
axguard contribute prepare

---

Responsible Use

AXguard is for authorized security testing and defense.

Only scan systems, applications, repositories, and infrastructure that you own or have permission to test.

---

Support

AXguard is free and open source.

If it helps you build safer software, a star on GitHub helps more builders find it.

You can also support the project and help fund more open-source security tools.

Website: awarexone.com
Email: [email protected] · [email protected] · [email protected]
Buy Me a Coffee: buymeacoffee.com/shuvonsec

Bitcoin	1GXwGqmLcnbZWgVNskUAZyw2cmqenkUFNY
Solana	4ArkPu1E7tkrt3d5X84grWzF1xjuLpScgGEy12Bp2cmE

---

License

MIT. See LICENSE.