Jacky CLI
jaswanthsai1/jacky-cli
Personal Hermes Agent fork with local or cloud models and a built-in bug-bounty hunt-loop methodology
Jacky CLI is a personalized fork of Nous Research's Hermes Agent that adds a bundled bug-bounty and offensive-security hunt-loop methodology. It runs against local models through Ollama or against any OpenAI-compatible or Anthropic-compatible cloud provider.
What Jacky CLI does
Jacky keeps the Hermes Agent feature set described in its README: a terminal interface with slash-command autocomplete, a tool-calling agent loop with more than 40 built-in tools, an MCP client, a gateway for Telegram, Discord, Slack, WhatsApp and Signal, a cron scheduler, subagents, agent-curated memory, autonomous skill creation and full-text session search. It supports six terminal backends: local, Docker, SSH, Singularity, Modal and Daytona.
The addition is a hunt-loop doctrine shipped under skills/: scope, recon, rank, enumerate, test, validate, chain and report, with finding-validation gates, evidence-hygiene rules and report-writing formulas. Models can be switched with the jacky model command, either to an Ollama model for offline work or to a cloud provider with your own key. A one-line installer clones the repository to ~/.jacky-cli, runs setup.sh and puts jacky on the PATH, and the tool can also be installed from npm or PyPI. A native Windows script exists.
Key features
- Local models through Ollama or any OpenAI- or Anthropic-compatible provider
- Bundled bug-bounty hunt-loop skills from scope to report
- Hermes-style learning loop, skills, memory and cron scheduler
- Telegram, Discord, Slack, WhatsApp and Signal gateway
- Six terminal backends including Docker, SSH, Modal and Daytona
- Install through a shell script, npm or pip
When to use it
- Running a bug-bounty workflow with a local model so nothing leaves the machine
- Switching between Ollama and a cloud provider with one command
- Starting from the Hermes Agent feature set with an offensive-security methodology preloaded
Who it is for: Security researchers and bug-bounty hunters who want a Hermes-based agent CLI with a built-in methodology.
How it fits with Hermes Agent
A fork of Hermes Agent with its own jacky command, not a plugin for an existing Hermes install.
How to install Jacky CLI
These commands are copied from the project's README. Check the repository for the latest steps before you run them.
curl -fsSL https://raw.githubusercontent.com/jaswanthsai1/jacky-cli/main/install.sh | bash
npm install -g jacky-cli-agent
pip install jacky-cliRequirements: Python 3.11 to 3.13 on Linux, macOS or Windows; optionally Ollama for local models, or an API key for a cloud provider
Note: It is a personal fork and not an official Nous Research release, and its license is reported as NOASSERTION.
FAQ
What is Jacky CLI?
Jacky CLI is an AI agent CLI forked from Nous Research's Hermes Agent. It supports local and cloud models and ships a bug-bounty hunt-loop methodology as skills.
Does Jacky CLI work with Hermes Agent?
It is a fork of Hermes Agent rather than an add-on. It has its own jacky command and keeps Hermes features such as skills, the gateway and the cron scheduler.
How do I install Jacky CLI?
On Linux or macOS, run the curl install script, which clones to ~/.jacky-cli and links jacky onto your PATH. You can also run npm install -g jacky-cli-agent or pip install jacky-cli.
Similar automations for Hermes Agent
All automationsHarper, a long-only virtual portfolio investing skill for Hermes Agent with an auditable SQLite ledger
MariyaSha Hermes Business AgentTutorial files for a 24/7 customer service agent built on Hermes, Telegram and Python
Microck Hermes Nightshift GLMOvernight bot for Hermes Agent cron that opens code quality pull requests and issues on your repos
Zuga-Technologies ZugaMindAlways-on attention sidecar that wakes Claude Code, OpenClaw, Codex or Hermes only when something matters
yassirboudda Hermes OSINT CTF AgentOSINT CTF agent template with username pivots, Reddit archives and swappable LLM providers
supermodeltools Big IronArchived prototype that ran an eight-phase SDLC in Hermes Agent, gated by a Supermodel code graph
Related guides: Build agent teams with the Hermes kanban board · How to run Hermes Agent securely