Hermes Agent v0.17.0 (v2026.6.19) — The Reach Release¶

Release Date: June 19, 2026 Since v0.16.0: ~1,475 commits · ~800 merged PRs · 1,693 files changed · 235,390 insertions · 50,730 deletions · 300+ issues closed · 245 community contributors

The Reach Release. v0.16.0 put Hermes on your desktop. v0.17.0 is about how far that reach extends — across new places to talk to it, deeper into the tools you already use, and smarter about how work gets done.

✨ Highlights¶

📱 iMessage Support via Photon (NEW)¶

No Mac relay required. Talk to Hermes through iMessage using Nous-managed Photon infrastructure. - hermes photon login — one-command setup - Blue-bubble delivery with full message fidelity - No Apple hardware dependency

🌊 Raft Agent Network Gateway (NEW)¶

Connect Hermes to Raft via wake-channel bridge — a privacy-preserving agent-to-agent network. - Cross-agent dispatch and coordination - Privacy-preserving connection model - Opt-in bridge architecture

🖥️ Desktop App Overhaul¶

Massive UX upgrade across the Electron desktop app: - Rebindable keyboard shortcuts panel — customize every action - Native OS notifications — system-level alerts when Hermes needs you - Live subagent watch-windows — see what your subagents are doing in real time - Composer model selector with per-model presets - VS Code theme support — install any Marketplace theme - Automatic RTL/bidi text direction - Resizable VS Code-themed terminal pane - Per-thread composer drafts

🤖 Background & Async Subagents¶

delegate_task(background=true) returns immediately — subagent runs in background, results come back when ready. No more blocking on long chains.

🖼️ Image Editing¶

The image_generate tool now supports editing and transforming existing images. Route edits to the appropriate endpoint automatically.

📋 Automation Blueprints (NEW)¶

Schedule recurring tasks via natural-language forms — no cron syntax required. "Every morning at 8 AM, check my calendar and summarize" just works.

🧠 Model & Provider Expansion¶

Cursor Composer model via xAI Grok — grok-composer-2.5-fast with full 200k context via OAuth
Full profile builder in dashboard — configure models, skills, MCP servers without touching config.yaml

🛒 Skills Hub Rehaul¶

Featured skills with curated discovery
Integrated security scans on every skill
Real browser-based browsing experience
Quality scoring and community ratings

🧩 Memory Tool Upgrade¶

Atomic batch operations (add/replace/remove) against a character budget. Smarter memory compaction with no loss of critical context.

🔐 Security & Dashboard¶

Hardened dashboard login — proper 401 handling, public_url warning
Promptware defense: Brainworm-class attack patterns blocked; memory scanned at load; tool outputs delimited
session_search rebuilt: 4,500× faster, zero LLM cost

📱 Messaging Platforms¶

Official WhatsApp Business Cloud API adapter — native Meta-hosted channel
Rich text for Telegram — Bot API 10.1 with native markup (on by default)

💰 Cost Optimization¶

Curator cost optimization — LLM consolidation pass now off by default; deterministic pruning is free
Token-aware tool dispatch reducing per-turn overhead

🖥️ Desktop App Details¶

UX & Productivity¶

Rebindable keyboard shortcuts panel — full customization
Live subagent watch-windows with streaming output
Composer model selector — switch models mid-conversation
Per-model presets for quick switching
Automatic RTL/bidi text direction detection
Resizable VS Code-themed terminal pane
Per-thread composer drafts (never lose an in-progress message)
Install any VS Code Marketplace theme

Notifications & Awareness¶

Native OS notification integration
Background needs-input indicators
Subagent completion alerts

🌐 Network & Connectivity¶

iMessage (Photon)¶

hermes photon login for setup
Managed infrastructure — no Mac required
Full message fidelity on blue bubbles

Raft Agent Network¶

Wake-channel bridge for agent-to-agent communication
Privacy-preserving connection model
Cross-agent task delegation

WhatsApp Business¶

Official Meta Cloud API adapter
Native business messaging support
Template management and approval flows

🏗️ Core Agent¶

Subagents¶

delegate_task(background=true) — async, non-blocking
Subagent results delivered when ready
Live watch-windows in desktop app

Tools¶

image_generate now supports image editing/transformation
session_search 4,500× faster, deterministic (no LLM cost)
Memory tool: atomic batch operations with character budget

Automation Blueprints¶

Natural-language scheduling — no cron syntax
Recurring task definitions with plain English
One-off and repeating patterns supported

📦 Skills & Ecosystem¶

Skills Hub 2.0¶

Featured skills with curated discovery
Per-skill security scanning
Browser-based exploration experience
Community ratings and quality scores

Profile Builder¶

Full model configuration from dashboard
MCP server management without editing config.yaml
Skill enable/disable from web interface

📱 Messaging Platforms¶

Platform	New in v0.17.0
iMessage	Full support via Photon — no Mac required
WhatsApp	Official Business Cloud API adapter
Telegram	Rich text via Bot API 10.1 (native markup, on by default)

🔒 Security¶

Promptware defense — Brainworm-class attack patterns blocked; memory scanned at load time; tool outputs delimited
Dashboard auth hardened — proper 401 handling, public_url exposure warning
session_search rebuilt as deterministic tool (no LLM → no prompt-injection surface)
Per-skill security scanning in Skills Hub

⚡ Performance¶

Improvement	Impact
`session_search` rebuild	4,500× faster, zero LLM cost
Curator deterministic pruning	Free consolidation, no LLM calls
Token-aware tool dispatch	Reduced per-turn overhead
Background subagents	Non-blocking multi-agent workflows

👥 Contributors¶

245 community contributors — the largest contributor cohort in Hermes Agent history.

*

This Hermes repo is one of the largest structured collections of public AI, automation, business, and technology documentation. Content remains attributed to original authors and repositories. Indexed and organized by www.CorpusIQ.io.