Violin ☤ — Supervised Agentic Hermes Pentest Profile

31 playbooks · 8 references · 0 plugins · 0 brokers · 100% Hermes built-in tools

Violin is a pluginless Hermes Agent profile for supervised, authorised penetration tests — from reconnaissance through safe exploit validation to reporting. It uses only Hermes' built-in toolsets, skill-based playbooks, and lightweight guard scripts. No custom plugin code, no broker service, no external orchestrator.

hermes profile install https://github.com/Dan-StrategicAutomation/violin
hermes -p violin

---

Features

🔬 31 Methodology Playbooks	7 phase playbooks (scoping → recon → exploitation → reporting) + 24 per-vulnerability-class playbooks covering OWASP Top 10, OWASP API Top 10, LLM Top 10, and beyond.
🛡️ Multi-Layer Safety	Interactive scoping (8 questions) → scope validation → machine guard check → user approval gates — every target-touching command is validated before execution.
🧠 Autonomous Tool Discovery	Detects installed tools, searches alternatives, reads docs, and asks to install what's missing. Never silently skips a step.
🌐 Browser + Web Research	Browser toolset for website enumeration (login forms, dashboards, DOM inspection). Web toolset for CVE lookup, exploit search, and OSINT.
📋 Evidence-Driven Reporting	Reproducible evidence with screenshots, tool output, request/response pairs. Structured CVSS 3.1 scoring with L1–L4 severity levels.
🔗 Hermes-Native	Inherits your existing Hermes provider/model. No extra API keys, no per-profile credentials, no lock-in.

---

Quick Start

# 1. Install the profile
hermes profile install https://github.com/Dan-StrategicAutomation/violin

# 2. Start a session
hermes -p violin

# 3. Let Violin ask 8 scoping questions, then run your test
> Run a pentest against example.com

Prerequisites

• Hermes Agent >= 0.18.0 — installed and on your PATH
• Hermes provider configured — Violin inherits your normal Hermes provider/model. No Violin-specific API key required.
• Kali Linux or Parrot OS recommended — Docker Kali, WSL, macOS, Windows, and remote jump boxes also work (the scoping phase adapts to your environment).

Set as default profile

hermes profile use violin

Now every hermes session loads Violin automatically.

---

Engagement Workflow

flowchart LR
    A["1. Scoping"] --> B["2. Recon"]
    B --> C["3. Vuln Research"]
    C --> D["4. Exploitation"]
    D --> E["5. Reporting"]
    E --> F["6. Retrospective"]
    
    A -.->|"clarify"| G("Approval Gate")
    B -.->|"guard check"| G
    C -.->|"guard check"| G
    D -.->|"clarify + guard"| G
    
    style A fill:#1a1a2e,stroke:#e94560,stroke-width:2px
    style B fill:#16213e,stroke:#0f3460,stroke-width:2px
    style C fill:#1a1a2e,stroke:#e94560,stroke-width:2px
    style D fill:#16213e,stroke:#0f3460,stroke-width:2px
    style E fill:#1a1a2e,stroke:#e94560,stroke-width:2px
    style F fill:#16213e,stroke:#0f3460,stroke-width:2px
    style G fill:#2d2d2d,stroke:#ffd700,stroke-width:2px

What each phase does

Phase	Action	Safety Gate
1. Scoping	8 questions via clarify — target type, testing mode, risk tolerance, authorisation	User approval
2. Reconnaissance	Passive OSINT → tech detection → active scanning	Guard + approval
3. Vuln Research	CVE lookup, exploit search, attack surface analysis	Guard check
4. Exploitation	Safe PoC validation per vulnerability class	Guard + user approval
5. Reporting	Evidence compilation, CVSS scoring, remediation	—
6. Retrospective	Gap analysis, playbook coverage update	Mandatory

---

Architecture

graph TB
    subgraph "Your Machine"
        HE["Hermes Agent"]
        VI["Violin Profile"]
        GUARD["violin_guard.py"]
    end
    
    subgraph "Violin Skills"
        SK["SKILL.md"]
        PB["31 Playbooks"]
        REF["8 References"]
        TEMP["2 Templates"]
    end
    
    subgraph "Hermes Built-in Tools"
        T["terminal"]
        W["web"]
        B["browser"]
        F["file"]
        CE["code_execution"]
        S["skills"]
        CL["clarify"]
        D["delegation"]
        V["vision"]
        TD["todo"]
    end
    
    LLM["Your LLM Provider"]
    
    HE -->|"hermes -p violin"| VI
    VI -->|"loads"| SK
    SK -->|"routes to"| PB
    PB -->|"uses"| T & W & B & F
    VI -->|"validates via"| GUARD
    HE -->|"calls"| T & W & B & F & CE & S & CL & D & V & TD
    HE -->|"inherits"| LLM
    
    style HE fill:#2d2d2d,stroke:#ffd700,stroke-width:2px
    style VI fill:#1a1a2e,stroke:#e94560,stroke-width:2px
    style GUARD fill:#16213e,stroke:#0f3460,stroke-width:2px

---

Toolsets

10 pentest-relevant Hermes toolsets enabled

#	Toolset	Purpose
1	terminal	Run pentest tools, discover tools, read docs
2	web	CVE research, exploit search, OSINT, tool discovery
3	browser	Website enumeration — forms, DOM, screenshots, link crawling
4	file	Evidence, reports, scope files
5	code_execution	Parse output, write custom exploit scripts
6	skills	Load pentest playbooks
7	todo	Track engagement phases and tasks
8	clarify	Ask user for approvals and decisions
9	delegation	Parallel recon and exploit development
10	vision	Screenshot analysis and visual evidence

Note: hermes tools --summary may show Hermes' internal kanban category even when no kanban_* tools are exposed. Violin's config.yaml intentionally enables only the 10 toolsets above.

Conversation & memory isolation

Violin is isolated by design:

• memory.memory_enabled: false — no global MEMORY.md recall/write
• memory.user_profile_enabled: false — no global USER.md access
• session_search is not enabled — cannot search unrelated chats
• Engagement continuity lives in project files (scope docs, evidence, notes, reports)
• Start a fresh Hermes conversation per engagement to keep boundaries clean

---

Safety Model

Multi-layer safety architecture

flowchart LR
    subgraph "Layer 1"
        A["8 Scoping Questions"]
        B["Written Authorisation"]
    end
    subgraph "Layer 2"
        C["violin_guard.py validate-scope"]
    end
    subgraph "Layer 3"
        D["violin_guard.py check-command"]
    end
    subgraph "Layer 4"
        E["clarify approval gate"]
    end
    subgraph "Layer 5"
        F["Standards & Blocked Actions<br/>in references/standards.md"]
    end
    
    A --> B --> C --> D --> E --> F

• Authorised testing only — no probing before scoping is complete and scope is approved
• Approval gates — scope, active recon, and exploitation each require explicit user approval via clarify
• Machine check — every target-touching command runs through violin_guard.py before execution (exit code 0=allowed, 1=blocked, 2=review required)
• Non-destructive by default — exploitation limited to safe, reproducible PoC
• Evidence-first — every finding backed by reproducible tool output, screenshots, request/response pairs
• Frameworks-grounded — every action justifiable under PTES, OWASP WSTG, or NIST SP 800-115

The full policy for approval tiers, blocked actions, evidence handling, rate limits, and scope allowlists lives in skills/pentest/references/standards.md.

Blocked actions

Action	Default Status
Credential attacks (bruteforce, spraying)	❌ Blocked — requires RoE carve-out
Social engineering (phishing, vishing)	❌ Blocked
Post-exploitation beyond proof	❌ Blocked
Persistence (backdoors, cron, services)	❌ Blocked
Stealth / evasion (anti-forensics, log tampering)	❌ Blocked
Malware deployment	❌ Blocked
Destructive payloads (DROP TABLE, rm -rf)	❌ Blocked
Data exfiltration	❌ Blocked — minimal PoC evidence only

---

Playbook Library

7 core phase playbooks

Phase	Playbook	Description
Scoping	playbooks/scoping.md	Scope definition, RoE, target validation
Reconnaissance	playbooks/recon.md	Passive & active recon, asset enumeration, tech detection
Vulnerability Research	playbooks/vuln-research.md	CVE lookup, manual analysis, tool scanning
Exploitation	playbooks/exploitation.md	PoC development, exploitation, evidence capture
Post-Exploitation	playbooks/post-exploitation.md	Limited: prove access extent, no lateral movement
Reporting	playbooks/reporting.md	Findings doc, risk rating, remediation, executive summary
Tools	playbooks/tools.md	Tool setup, discovery, and management

24 per-vulnerability-class playbooks

Vulnerability Class	Playbook	OWASP Mapping
SQL Injection (SQLi)	playbooks/sqli.md	A03:2021 — Injection
Cross-Site Scripting (XSS)	playbooks/xss.md	A03:2021 — Injection
Command Injection	playbooks/command-injection.md	A03:2021 — Injection
SSRF	playbooks/ssrf.md	A10:2021 — SSRF
Server-Side Template Injection	playbooks/ssti.md	A03:2021 — Injection
Path Traversal	playbooks/path-traversal.md	A01:2021 — Broken Access Control
IDOR / Access Control	playbooks/idor-access-control.md	A01:2021
Authentication Bypass	playbooks/auth-bypass.md	A07:2021 — Auth Failures
JWT Attacks	playbooks/jwt-attacks.md	A07:2021 — Auth Failures
Insecure Deserialization	playbooks/deserialization.md	A08:2021 — Integrity Failures
XML External Entities (XXE)	playbooks/xxe.md	A05:2021 — Security Misconfig
NoSQL Injection	playbooks/nosql-injection.md	A03:2021 — Injection
LLM Prompt Injection	playbooks/llm-prompt-injection.md	OWASP LLM Top 10
Business Logic Flaws	playbooks/business-logic.md	A04:2021 — Insecure Design
API Security	playbooks/api-security.md	OWASP API Top 10
Input Validation	playbooks/input-validation.md	A04:2021 — Insecure Design
Cryptographic Issues	playbooks/cryptographic-issues.md	A02:2021 — Crypto Failures
Security Misconfiguration	playbooks/security-misconfiguration.md	A05:2021 — Security Misconfig
Security Through Obscurity	playbooks/security-through-obscurity.md	A05:2021 — Security Misconfig
Unvalidated Redirects	playbooks/redirects-unvalidated.md	A01:2021 — Broken Access Control
Anti-Automation	playbooks/anti-automation.md	OWASP Automated Threats
Observability Failures	playbooks/observability-failures.md	A09:2021 — Logging Failures
Supply Chain	playbooks/supply-chain.md	A06:2021 — Vulnerable Components
CSRF	playbooks/csrf.md	A01:2021 — Broken Access Control

8 reference documents

Reference	Path
Methodology (PTES/OWASP/NIST)	references/methodology.md
CVE Research APIs	references/cve-apis.md
Tool Discovery	references/tool-discovery.md
Tool Catalog	references/tool-catalog.md
Kali/Parrot Tool Paths	references/kali-parrot-paths.md
Operational Standards	references/standards.md
Retrospective	references/retrospective.md
Real-World Coverage Matrix	references/real-world-coverage-matrix-template.yaml

2 templates

Template	Path
Scope Document (YAML)	templates/scope-template.yaml
Pentest Report (Markdown)	templates/report-template.md

---

Repository Layout

violin/
├── .hermes.md              # Project-level agent context
├── SOUL.md                 # Agent identity — senior pentester persona
├── config.yaml             # Profile config (toolsets, safety, memory)
├── distribution.yaml       # Hermes distribution manifest
├── LICENSE                 # MIT License
├── README.md               # This file
├── assets/                 # Brand assets (logo)
├── scripts/                # Guard & smoke tests
│   ├── violin_guard.py     # Scope, command & release validation
│   ├── smoke-test.sh       # Kali/Parrot release smoke
│   ├── smoke-test.ps1      # Windows supplemental smoke
│   └── kali.sh             # Docker Kali helper
└── skills/pentest/         # Pentest methodology
    ├── SKILL.md            # Orchestrator skill
    ├── playbooks/          # 31 playbooks
    ├── references/         # 8 reference documents
    └── templates/          # 2 templates (scope, report)

---

Release Verification

# Run all release checks
python scripts/violin_guard.py check-release

# Or use the smoke tests
./scripts/smoke-test.sh --no-install          # Linux/macOS
.\scripts\smoke-test.ps1 -NoHermes            # Windows

The guard check validates: YAML structure, 31 playbooks present, required sections present, no stale evidence paths, all markdown references resolve, distribution manifest covers all owned files.

---

Optional: Kali Docker Container

One-time setup for a full Kali toolchain on any OS

# Create the container
docker pull kalilinux/kali-rolling
docker create -it --name kali-pentest \
  -v /path/to/violin/engagements:/engagements \
  kalilinux/kali-rolling bash
docker start kali-pentest

# Install tools (~600 packages, 5-10 min)
docker exec kali-pentest apt update
docker exec kali-pentest apt install -y kali-linux-headless

# Verify
docker exec kali-pentest nmap --version

Use scripts/kali.sh as shorthand: kali nmap -sV target.com instead of docker exec kali-pentest nmap -sV target.com.

---

Optional MCP Servers

Enhance Violin with community MCP servers

MCP Server	Install	Purpose
onlinecybertools-mcp-server	npx -y onlinecybertools-mcp-server	280+ OSINT/recon tools — whois, SSL check, JWT decode, network diag, hashes
runyourempire/4DA	npx -y @runyourempire/4DA	Live CVE scanning, dependency health, upgrade planning

Add them to config.yaml under mcp_servers: after install. See Hermes MCP docs.

---

Contributing

See CONTRIBUTING.md for development setup, PR process, and code style.

Security

See SECURITY.md for reporting vulnerabilities.

---

License

MIT — see LICENSE.