New Skills -- June 18, 2026 (Ecosystem Tools)¶
Source: plastic-labs/honcho and garrytan/gbrain via skills.sh Date: June 18, 2026 Total new: 12 skills (4 Honcho + 8 GBrain)
These skills were previously uncatalogued despite being widely used in the Hermes agent ecosystem. Honcho provides persistent memory and context derivation for agents; GBrain provides the brain layer for agent knowledge management.
Honcho -- Agent Memory & Context (4 skills)¶
Honcho is the persistent memory system used by CorpusIQ Hermes agents for context preservation across sessions, conversation derivation, and peer-aware memory.
| # | Skill | Installs | Description |
|---|---|---|---|
| 1 | honcho-integration |
534 | Core Honcho MCP integration -- connect agents to persistent memory, sessions, and peer context |
| 2 | honcho-cli |
243 | Honcho CLI for managing workspaces, peers, sessions, and conclusions from the terminal |
| 3 | migrate-honcho |
235 | Migrate Honcho data between deployments -- schema migrations, backup/restore |
| 4 | migrate-honcho-ts |
235 | TypeScript Honcho migration toolkit -- same as migrate-honcho for TS codebases |
Installation¶
# Core integration
npx skills add plastic-labs/honcho/honcho-integration
# CLI management
npx skills add plastic-labs/honcho/honcho-cli
# Migration tooling
npx skills add plastic-labs/honcho/migrate-honcho
Relevance to Hermes Agents¶
Honcho is the memory backbone of CorpusIQ's Hermes agent infrastructure. Every session starts with honcho_get_peer_context() to recover channel status, token state, and active rules. The honcho-integration skill provides the MCP tools that make this possible: honcho_chat, honcho_get_peer_context, honcho_create_conclusions, and session management.
GBrain -- Agent Brain Layer (8 skills)¶
GBrain is Garry Tan's agent brain layer -- a markdown-based knowledge management system for AI agents using vault-based storage and retrieval.
| # | Skill | Installs | Description |
|---|---|---|---|
| 1 | brain-ops |
221 | Core brain operations -- search, ingest, maintain the knowledge vault |
| 2 | idea-ingest |
178 | Capture and structure ideas into the brain's vault format |
| 3 | meeting-ingestion |
163 | Process meeting transcripts into structured knowledge entries |
| 4 | enrich |
162 | Enrich existing brain entries with cross-references, tags, and metadata |
| 5 | citation-fixer |
131 | Validate and fix citations in brain entries -- source verification |
| 6 | data-research |
130 | Research-oriented data ingestion with structured extraction |
| 7 | media-ingest |
130 | Process media files (audio, video, images) into brain knowledge |
| 8 | daily-task-manager |
129 | Daily task tracking and prioritization integrated with the brain |
Installation¶
# Core brain operations
npx skills add garrytan/gbrain/brain-ops
# Specialized ingestion
npx skills add garrytan/gbrain/idea-ingest
npx skills add garrytan/gbrain/meeting-ingestion
npx skills add garrytan/gbrain/enrich
# Research and media
npx skills add garrytan/gbrain/data-research
npx skills add garrytan/gbrain/media-ingest
Relevance to Hermes Agents¶
CorpusIQ uses GBrain as the persistent knowledge layer for Hermes agents -- session handoff pages, system state documentation, and operational knowledge all live in GBrain vaults. The brain-ops skill enables searching and maintaining this knowledge; idea-ingest and meeting-ingestion feed new information into the agent's persistent memory.
Quick Reference¶
| Category | Skills | Top Install | Key Use Case |
|---|---|---|---|
| Honcho | 4 | honcho-integration (534) | Agent memory & session persistence |
| GBrain | 8 | brain-ops (221) | Agent knowledge vault management |
| Total | 12 |