Implemented in the current build.
The agent sits on top of the scope and consent layer, not beside it, so the model can propose a step but cannot approve its own scope or run anything intrusive on its own.
- interactive agentimplemented
- A full-screen terminal session you talk to. Conversational TUI with a fixed input box, a scrolling log of the agent's reasoning and tool activity, an animated thinking status, live steering while it works, cancel with Esc, slash commands and input history. It answers you and only scans when you ask.
- pwner chat
- tool-calling agentimplemented
- Bring-your-own-model planning. The agent reasons, calls a tool, reads the result and iterates, then reports. Runs on OpenAI, DeepSeek, Anthropic, Gemini, or a local Ollama, Lemonade or llama.cpp server, or a deterministic suite with no model at all.
- --no-ai
- commandsimplemented
- A handful of commands, no config files. First run sets your provider and key once; after that you open a session or run a one-shot scan. Change the model later with
/modelin a session, orpwner --pickat launch. - pwner · chat · scan · pci · setup
- read-only toolsimplemented
- Discovery, fingerprinting and detection. ping_sweep, dns_enum, port_scan, service_detect, os_detect, banner_grab, http_probe, vuln_scan, nuclei_scan, exploit_search, ftp_anon and snmp_check, all scope-locked. Brute-force is consent-gated.
- wraps nmap + nuclei
- safety layerimplemented
- Scope lock and consent gates outside the model. Every tool refuses hosts outside the target you named, intrusive actions are denied by default, and the raw-nmap escape hatch blocks intrusive scripts.
- --auto-consent
- PCI segmentationimplemented
- 11.4.5 / 11.4.6 from an out-of-scope segment. Tries to reach the CDE and reports whether the boundary held, with no model involved. A closed port answering with a RST counts as a failure of isolation, not a pass; a host RST cannot be told from a firewall reject over TCP, so that case is returned as inconclusive rather than guessed. Deterministic, so re-running it after a fix is the 11.4.4 retest. It does not make anyone compliant.
- pwner pci
- outputimplemented
- Findings and a live map. Confirmed findings with severity, evidence and remediation, printed in the terminal or as JSON, an ephemeral topology tree, and CIDR scanning that sweeps, scans hosts in parallel and runs nuclei over what it finds.
- --json
Deeper, not wider.
Next is mostly about detecting more without loosening the scope and consent model, and about making what the agent finds easier to hand off.
- More known-issue templates and more anonymous and weak-access checks, each surfaced as a finding with evidence, not a recipe.
More detection coverage
more nuclei / NSE (planned) - Turn a session's findings into a shareable Markdown or HTML engagement report, built from the same severity, evidence and remediation fields.
Report export
pwner report (planned) - Broader support for tool-calling models, hosted and local, so you can run the agent on whatever you already trust with the data.
More providers
--provider (expanding) - Decorating an async function with
Tool API
@tool (shipped)@toolregisters it, advertises it to the model, makes it dispatchable and scope-guards it — one declaration site, and the guard is applied for you rather than by you. See where it plugs in.
- coveragenmap + nuclei, more coverage, auth'd scans
- reportingfindings + JSON, report export, shared reports
- extensibilityPython tools, tool API, tool packs
Why the tool API comes with guardrails. A new tool is the one place added code could reach a host it should not. It registers behind the same scope lock and consent gate as the built-in tools, or it does not register at all.
Maybe, if the core holds up.
These depend on what people actually need once the tool API is in their hands. None of them change the safety model.
- Authenticated scan modulesRead-only checks that use credentials you supply, held to the same scope lock and consent gates as everything else.
- Multi-operator engagementsSeveral testers share one findings report, each with their own consent decisions.
- Signed, reproducible buildsRelease artifacts you can rebuild and compare yourself.
- Your ideaOpen an issue in the project repository once it is public, and say what you would test.
Release notes will appear here once there is a release. No versions or dates are set.
No roadmap item ships if it makes the scope lock optional or lets the agent approve its own consent gates.
Features are reviewed against four things that do not change: the scope lock, the consent gates, the read-only default, and authorized-testing-only use. A pull request that weakens one of them will not be merged, however useful it is.
scope lock firstNew tools go through the same scope check as the built-in ones, outside the model.consent stays with youNo feature lets the agent self-approve an intrusive action.read-only firstNew detection confirms a weakness read-only; anything that exploits it stays behind a consent gate.your model, your machineNo telemetry. Only the model endpoint you choose ever sees data.How to contribute.
Small, reviewed changes beat large surprises. Most first contributions are docs, error messages and new read-only checks.
- Pick an issue from the list, or open one describing what you want to change.
- Say you are on it in a comment so two people do not build the same thing.
- Test in a lab. Run against your own lab range only, such as 10.0.0.0/24, never someone else's network.
- Open a pull request with a note on how the change interacts with the scope lock and consent gates.
- Respond to review. Reviewers will reply, and so should you.
Good first issues (suggestions)
- [core]Add a provider preset for another tool-calling model
- [core]Improve the error text when nmap or nuclei is missing
- [docs]Walk through adding a new read-only tool end to end
- [detect]Add a nuclei template group for a common service
These are suggestions, not filed tickets — the tracker is open and empty, so take one or bring your own.
- GitHub
- Issues
- Contributing guide
- Chat none yet
There is no chat room yet. Bugs and design questions go to GitHub issues; a vulnerability in pwner itself goes to private advisories instead of a public issue.