Back to Index
Tools
30+ tool functions across 8 categories
File Operations
read_file
Read text files with line numbers and pagination. Auto-extracts .ipynb, .docx, .xlsx.
write_file
Write content to a file, completely replacing existing content. Creates parent directories.
patch
Targeted find-and-replace edits. Fuzzy matching with 9 strategies.
search_files
Ripgrep-backed search. Content search (regex) or file search (glob).
Terminal & Execution
terminal
Execute shell commands in bash. Persistent state. Foreground (instant return) or background (long-lived).
process
Manage background processes — list, poll, log, wait, kill, write stdin.
read_terminal
Read what's shown in the Hermes desktop GUI terminal pane.
execute_code
Run Python scripts that call Hermes tools programmatically. Loops, conditionals, processing.
Web & Research
web_search
Search the web. Supports operators: site:, filetype:, intitle:, -term, "exact phrase".
web_extract
Extract page content as markdown. Works with PDFs (arxiv papers, documents).
browser_navigate
Navigate to a URL. Returns compact page snapshot with interactive elements and ref IDs.
browser_click · browser_type · browser_scroll · browser_press · browser_back
Full browser interaction suite. Click, type, scroll, keyboard, navigation.
browser_snapshot
Text-based snapshot of accessibility tree. Compact or full modes.
browser_console
Get console output and JS errors. Evaluate JavaScript in page context.
browser_vision
Screenshot the page for visual inspection. Can annotate interactive elements.
browser_get_images
List all images on the page with URLs and alt text.
Media
image_generate
Text-to-image and image-to-image via FAL.ai · FLUX 2 Klein 9B. Landscape, square, portrait.
vision_analyze
Load images into context for visual inspection. URL, local path, or data URL.
video_analyze
Analyze video (mp4, webm, mov, avi, mkv). Multimodal model understanding.
video_generate
Text-to-video or image-to-video generation. No backend configured.
text_to_speech
Convert text to speech. ElevenLabs backend. Renders as voice bubble on Telegram.
Memory & Sessions
memory
Save durable facts to persistent memory. Injected into every future turn.
session_search
FTS5 search across past session transcripts. Discover, scroll, read, browse.
skill_view · skills_list
Load and list skills (procedural memory). 43 installed.
skill_manage
Create, update, delete skills. Full lifecycle management.
Delegation & Automation
delegate_task
Spawn subagents for isolated tasks. Single or batch (up to 3 parallel).
cronjob
Schedule recurring jobs. Create, list, update, pause, resume, run, remove.
todo
Session task list. Complex tasks with 3+ steps.
clarify
Ask the user for clarification. Multiple choice or open-ended.
mixture_of_agents
Route hard problems through 4 reference models + 1 aggregator. Use sparingly.
Integrations
MCP: Linear
Full Linear API — issues, projects, initiatives, teams, cycles, documents, comments, attachments. OAuth pending.
GitHub CLI (gh)
Repos, issues, PRs, code review, CI/CD. Authenticated as F1SOURCE.
Vercel CLI
Deploy, link, manage domains. v54.17.1. Connected to frankgomez4-1952.
Computer Use
computer_use (via skill)
Drive user's desktop — clicking, typing, scrolling, dragging. Cross-platform. Background operation.
Total tool functions
30+
Categories
8
Subscription tools
Browser Use · FAL.ai · ElevenLabs · Groq Whisper
MCP tools
Linear (OAuth pending)
Tools · SourceMEM.com · 30+ functions