v0.2.6
Armur-Ai/Pentest-Swarm-AIv0.2.6Sep 13, 2026by github-actions[bot]
Upgrading?
AI Summary
This release introduces a comprehensive dashboard overhaul featuring live progress visualization, attack graph tracking, and an end-of-run report. Key additions include a killswitch, spend cap controls, and a fan-out mesh visualization for exploit probes. The engine also gains per-task model routing capabilities, while the TUI is updated with clearer progress indicators.
Key Highlights
- Dashboard Overhaul: Live progress, attack graph, and inline findings
- Fan-out Mesh Visualization: Visualizing exploit probe spreading
- Spend Cap & Killswitch: Financial controls to halt runs
- Per-task Model Routing: Mixture of AI models based on tasks
- TUI Improvements: Clearer progress counters and live activity
New Features
- Live progress console and dashboard visualization
- Attack graph growing toward recon/exploit
- Inline findings expansion (replacing floating modal)
- End-of-run report with proven attack chains
- Killswitch to stop running swarms
- Exploit probe fan-out mesh visualization
- USD spend cap for swarm runs
- Per-task model routing for AI models
- Probe telemetry and event stream integration
- TUI activity line and live counters
Full Release Notes
## Pentest Swarm AI v0.2.6 Install via Homebrew: ``` brew install Armur-Ai/tap/pentestswarm ``` Install via AUR (once submitted): ``` yay -S pentestswarm-bin ``` Run via Docker (16 security tools pre-installed): ``` docker run --rm ghcr.io/armur-ai/pentestswarm:v0.2.6 scan example.com ``` ## Changelog ### Features * 14d56b42a336b86ab323c9350c87d345e14ce871: Merge feat(dashboard): live progress, growing attack graph, inline findings + report (@AkhilSharma90) * 0969df025a2d510effba9fc39f82dd123cf7a3e0: Merge feat: swarm fan-out mesh — exploit probe visualization (@AkhilSharma90) * 7f27ae49631c403857a6b0f6b30e4ffd68a3b08c: feat(cli): --budget spend cap + launcher 'Spend cap' field + dashboard killswitch wiring (@AkhilSharma90) * 99364f1653b8edc5f3d7c81d48f04e4d36f29b06: feat(dashboard): attack graph grows toward recon/exploit as events land (@AkhilSharma90) * 24cd9bcaf28fecd559e621e10f6340fdd0284961: feat(dashboard): findings expand inline instead of a floating modal (@AkhilSharma90) * 583e1fec2278f419bf65c31d607c39f18d9adacf: feat(dashboard): kickass end-of-run report + proven attack chains (@AkhilSharma90) * 4961a174645b66a9f7beb64c7545e16cb164d383: feat(dashboard): killswitch — Stop button posts /api/stop to halt the run (@AkhilSharma90) * 1fe46f3978bd24cf9f82a9bdab811eb2d1d31249: feat(dashboard): live at :7777 while the launcher is open + clearer chain status (@AkhilSharma90) * f974718416dc6d2d076b5dad1da4942dd1f2ae38: feat(dashboard): prominent live progress console below phase strip (@AkhilSharma90) * d4767f19af1adfc76ac2847bbd5c5abd9b71c4c7: feat(dashboard,tui): render the exploit probe fan-out mesh (@AkhilSharma90) * bcf7a9be5bfaea2e8064f7b72017f44be48b581b: feat(engine): hard USD spend cap + manual killswitch for swarm runs (@AkhilSharma90) * d7aaa680f02ccf88eb24ec4ff0985bdf10f12b2e: feat(engine): per-task model routing — a real 'mixture by task' on Together (@AkhilSharma90) * ac4602598ebf3c5f10c3f83dc44752d2373c25d4: feat(swarm): probe telemetry — EventProbe + SweepBOLAWithProbe callback (@AkhilSharma90) * 425b4b2303f28f68da1596779f979cfab2fb4273: feat(swarm): thread a probe sink from the exploit agent to the event stream (@AkhilSharma90) * 8d47fb588723f5746e0ed9b791387ea70764ddbc: feat(tui): clearer progress — live 'NOW ▸' activity line + counters (@AkhilSharma90) ### Docs * d576a87df61e68672c8e131641e3c2fef4083a66: docs(tui): launcher Together note reflects the real per-task model mixture (@AkhilSharma90) ### Other * 69ef1cba8b39789951ce86549e202918d7d757d4: chore(release): 0.2.6 — fan-out mesh, budget cap + killswitch, per-task routing, dashboard overhaul (@AkhilSharma90) --- Full source: github.com/Armur-Ai/Pentest-Swarm-AI
Related Repositories
Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
A fast reverse proxy to help you expose a local server behind a NAT or firewall to the internet.
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs
The world’s fastest framework for building websites.
Persistent memory system for AI coding agents. Agent-agnostic Go binary with SQLite + FTS5, MCP server, HTTP API, CLI, and TUI.