hermes-okf 0.5.5

Universal OKF-based memory system for Hermes agent — structured, persistent, agent-readable knowledge storage.

Hermes OKF (Open Knowledge Format) — Universal Memory for AI Agents

The first open-source memory system built on Google's Open Knowledge Format (OKF) for the Hermes agent ecosystem. pip install hermes-okf && hermes-okf install-plugin — two commands and you're live. The install command auto-configures ~/.hermes/config.yaml so hermes memory setup finds the plugin immediately. hermes okf search|list|show|snapshot|restore work out of the box.

Hermes OKF gives your AI agent a persistent, structured, version-controlled memory — no database, no lock-in, just markdown + YAML on your filesystem. Every decision, observation, and project context lives in a human-readable knowledge graph that your agent can read, write, and traverse programmatically.

---

Table of Contents

• Quick Start: Install as Hermes Plugin
• Hermes Plugin CLI Commands
• Standalone CLI
• Why OKF?
• Architecture
• Agent Integration
• RAG Integration (Optional)
• Development
• Troubleshooting
• Roadmap

---

Quick Start: Install as Hermes Plugin

Step 1 — Install from PyPI

pip install hermes-okf

Step 2 — Register the plugin in Hermes

Hermes discovers plugins from the ~/.hermes/plugins/ directory. Run the install command to create the plugin wrapper:

hermes-okf install-plugin

Expected output:

Installed hermes-okf plugin to /home/username/.hermes/plugins/hermes-okf
  Updated ~/.hermes/config.yaml

What this does: Creates ~/.hermes/plugins/hermes-okf/ and auto-updates ~/.hermes/config.yaml to add hermes-okf to plugins.enabled and set memory.provider. Hermes finds the plugin on next startup.

Step 3 — Activate the memory provider

Start Hermes and select hermes-okf from the interactive setup:

hermes memory setup

Select hermes-okf from the list

Uninstall

To remove the plugin wrapper from Hermes:

hermes-okf uninstall-plugin

This removes ~/.hermes/plugins/hermes-okf/ but does not delete your OKF bundle.

---

Hermes Plugin CLI Commands

When installed as a Hermes plugin, these subcommands are available under hermes okf:

# Search your OKF memory
hermes okf search "dark mode"

# List stored concepts (optionally filter by type)
hermes okf list --type Decision

# Show full content of a specific concept
hermes okf show config/agent
hermes okf show sessions/2026-06-14T22-14-58Z
hermes okf show sessions/2026-06-14T22-14-58Z --raw  # metadata stripped

# Save a snapshot
hermes okf snapshot --note "Before deployment"

# Restore from last snapshot
hermes okf restore

The show command displays the full concept with YAML frontmatter and markdown body, making it easy to inspect any piece of your agent's memory.

---

Standalone CLI

Even without the Hermes plugin, you can use hermes-okf as a standalone knowledge management tool:

# Initialise a new OKF bundle
hermes-okf init ./knowledge

# Validate conformance
hermes-okf validate --path ./knowledge

# List concepts
hermes-okf list --path ./knowledge

# Show a concept
hermes-okf show --path ./knowledge projects/my_project

# Search
hermes-okf search --path ./knowledge "ffmpeg GPU"

# View log
hermes-okf log --path ./knowledge

# Append to log
hermes-okf log-append --path ./knowledge "New decision made" --category Decision

# Graph inspection
hermes-okf graph-edges --path ./knowledge
hermes-okf graph-neighbors --path ./knowledge projects/my_project

# Save snapshot
hermes-okf snapshot --path ./knowledge --note "Before deploy"

# Build LLM context
hermes-okf context --path ./knowledge "What should I prioritize?"

# List sessions, plans, tools
hermes-okf sessions --path ./knowledge
hermes-okf plans --path ./knowledge
hermes-okf tools --path ./knowledge

---

Why Hermes OKF?

Feature	What You Get
🧠 Agent Memory	Persistent decisions, observations, and tool-call history across sessions
🔗 Knowledge Graph	Implicit graph from markdown links — no RDF, no Cypher
📁 Filesystem-First	Plain .md + YAML. cat it, grep it, Git it.
⚡ Zero-DB Core	Single dependency: pyyaml. Optional RAG via LangChain/ChromaDB.
🔌 Hermes Plugin	HermesOKFMemoryProvider — native MemoryProvider ABC, discovered via hermes-okf install-plugin
🎁 Hermes-Ready	Drop-in decorators: @memorize_decision, @memorize_tool
🔄 Resume	Stop and restart — the agent restores from its OKF bundle
📦 Portable	Clone a bundle to another machine — the agent resumes instantly.

---

Why OKF?

OKF (Open Knowledge Format) is a vendor-neutral, open specification published by Google Cloud on June 12, 2026. It formalizes the "LLM wiki" pattern into a portable standard: every concept is a .md file with YAML frontmatter, and markdown links create a knowledge graph.

"OKF is a vendor-neutral, agent- and human-friendly standard for representing the metadata, context, and curated knowledge that modern AI systems need." — Sam McVeety & Amir Hormati, Google Cloud

Why hermes-okf chose OKF:

OKF Principle	What it means for agents
Minimally opinionated	Only one required field: type. Everything else is up to the producer.
Producer/consumer independence	A human can write a bundle; an AI agent can read it. No lock-in.
Format, not platform	No proprietary runtime, no SDK, no cloud required. Just markdown files.
Human-readable	cat any file and understand it. Git diffs work out of the box.

References:

• Google Cloud Blog: Introducing the Open Knowledge Format
• OKF Specification (GitHub)
• Google Cloud Knowledge Catalog

---

Architecture

┌─────────────────────────────────────────────────────────────┐
│  HUMAN INTERFACE                                              │
│  ├─ hermes okf search|list|show|snapshot|restore  (Hermes CLI) │
│  ├─ hermes-okf init|validate|search|show...     (Standalone)  │
│  ├─ hermes-okf install-plugin / uninstall-plugin  (Plugin mgmt) │
├─────────────────────────────────────────────────────────────┤
│  HERMES PLUGIN LAYER                                          │
│  ├─ HermesOKFMemoryProvider  ← MemoryProvider ABC implementation│
│  ├─ plugin.py / cli_extension.py  ← CLI registration bridge    │
│  ├─ install_plugin.py  ← Creates ~/.hermes/plugins/hermes-okf/│
├─────────────────────────────────────────────────────────────┤
│  UNIVERSAL PROVIDER                                           │
│  ├─ HermesOKFProvider  ← Any Hermes agent can use it          │
│  ├─ HermesAgent / MemoryMixin  ← Drop-in decorators           │
│  ├─ HotMemoryBuffer  ← In-process fast write buffer           │
├─────────────────────────────────────────────────────────────┤
│  CORE OKF LAYER                                               │
│  ├─ OKFBundle  ← File I/O, concept CRUD, logging            │
│  ├─ Concept  ← Dataclass: type, title, body, metadata         │
│  ├─ GraphExtractor  ← Link traversal, tag clustering          │
│  ├─ SearchIndex  ← Full-text + fuzzy search                 │
│  └─ OKFValidator  ← Conformance checking                      │
├─────────────────────────────────────────────────────────────┤
│  PERSISTENCE                                                  │
│  └─ Filesystem (markdown + YAML frontmatter)                  │
└─────────────────────────────────────────────────────────────┘

Read the full architecture in docs/ARCHITECTURE.md.

---

Agent Integration (Memory Mixin)

For most Hermes users, the plugin approach above is recommended. The decorators below are for advanced use cases or custom agent classes.

from hermes_okf.agent import HermesMemoryMixin

class MyAgent(HermesMemoryMixin):
    def __init__(self):
        super().__init__("./agent_knowledge", agent_id="my-agent-v1")
        self.start_session()

    @HermesMemoryMixin.memorize_decision
    def choose_model(self, task: str) -> str:
        if "code" in task.lower():
            return "anthropic/claude-3.5-sonnet"
        return "openai/gpt-4o"

    @HermesMemoryMixin.memorize_tool
    def scrape_data(self, url: str) -> dict:
        return {"url": url, "items": 42}

# Run it
agent = MyAgent()
agent.choose_model("Write a Python script")
agent.scrape_data("https://example.com")

# Recall relevant context
context = agent.with_context("python script", top_k=3)

For the full HermesAgent (state-as-bundle), see docs/HERMES_INTEGRATION.md. For the universal provider usage, see docs/HERMES_USERS.md.

---

RAG Integration (Optional)

pip install hermes-okf[rag]

Feed your OKF bundle into LangChain + ChromaDB for vector retrieval:

from langchain_community.document_loaders import DirectoryLoader, TextLoader
from langchain.text_splitter import MarkdownHeaderTextSplitter
from langchain_chroma import Chroma
from langchain_openai import OpenAIEmbeddings

from hermes_okf.bundle import OKFBundle

bundle = OKFBundle("./my_knowledge")
loader = DirectoryLoader(
    str(bundle.root),
    glob="**/*.md",
    loader_cls=TextLoader,
    loader_kwargs={"encoding": "utf-8"},
)
docs = loader.load()

splitter = MarkdownHeaderTextSplitter(
    headers_to_split_on=[("#", "Header 1"), ("##", "Header 2")]
)
splits = [chunk for doc in docs for chunk in splitter.split_text(doc.page_content)]

vectorstore = Chroma.from_documents(
    documents=splits,
    embedding=OpenAIEmbeddings(),
    persist_directory="./chroma_okf_db",
)

# Query
retriever = vectorstore.as_retriever(search_kwargs={"k": 5})
results = retriever.invoke("What GPU decisions did we make?")

See examples/rag_integration.py for a complete example.

---

Development

git clone https://github.com/EliaszDev/hermes-okf.git
cd hermes-okf
pip install -e ".[dev]"
pytest

See CONTRIBUTING.md for full guidelines.

---

Troubleshooting

hermes-okf install-plugin fails

If the command is not found, make sure hermes-okf is installed:

pip install --upgrade hermes-okf
hermes-okf --version
# Expected: 0.5.0+

If hermes-okf itself is not in PATH, use the module form:

python -m hermes_okf.cli install-plugin

hermes memory setup doesn't show hermes-okf

1. Make sure you ran hermes-okf install-plugin (creates ~/.hermes/plugins/hermes-okf/)
2. Check the plugin directory exists:

   ls ~/.hermes/plugins/hermes-okf/
   # Should show: __init__.py  plugin.yaml
   

3. Ensure plugins.enabled in ~/.hermes/config.yaml is a YAML list:

   plugins:
     enabled:
       - hermes-okf
   

hermes okf show shows wrong model

As of v0.3.7, HermesOKFMemoryProvider.initialize() reads the actual Hermes model from config.yaml (top-level model or llm.model) and syncs it into the OKF config/agent concept. If you see an old model, restart your Hermes session to trigger re-initialization.

OKF bundle not found

hermes-okf init ~/.hermes/okf_memory

Windows: filename errors

Hermes-OKF uses Windows-safe filenames (2026-06-15T10-30-00Z instead of 2026-06-15T10:30:00Z). If you see errors with older versions, upgrade:

pip install --upgrade hermes-okf

---

What's New in v0.5.0

Critical memory provider integration fix. Previous versions shipped the OKF CLI and plugin skeleton, but the agent-side integration was broken — handle_tool_call() raised NotImplementedError, queue_prefetch() was a no-op, and system_prompt_block() didn't instruct the agent to use memory tools. Users had to manually edit their system prompt ("soul md") to get data retrieval working.

Fixed in v0.5.0

Component	Before (v0.4.x)	After (v0.5.0)
handle_tool_call()	❌ Not implemented → NotImplementedError	✅ Dispatches search_memory + snapshot_memory with JSON results
queue_prefetch()	❌ No-op → no background warming	✅ Background thread prefetch warming
system_prompt_block()	Passive ("inspect memory at...")	Active ("memories are auto-injected, use search_memory to...")
prefetch()	No error handling → crashes on broken bundles	✅ try/except with graceful fallback
Tool descriptions	Generic	Descriptive ("Use when you need to recall past decisions...")

What this means: A fresh pip install hermes-okf && hermes-okf-install now works out of the box. The agent gets told about auto-injected memories, can call search_memory/snapshot_memory tools without crashing, and prefetch is warmed in the background.

Upgrading

pip install --upgrade hermes-okf

No configuration changes needed — the plugin automatically uses the new integration.

---

Roadmap

#	Feature	Status	Priority	Notes
1	Hermes plugin (HermesOKFMemoryProvider)	✅ Shipped	P0	Full MemoryProvider ABC; hermes memory setup integration
2	Plugin installer (hermes-okf install-plugin)	✅ Shipped	P0	One-command registration in ~/.hermes/plugins/
3	Unified CLI (hermes-okf single entry point)	✅ Shipped	P0	install-plugin/uninstall-plugin subcommands replace standalone scripts
3	Universal provider (HermesOKFProvider)	✅ Shipped	P0	Any Hermes agent can use it out of the box
4	Two-memory model (hot + cold archive)	✅ Shipped	P0	Automatic flushing from hot buffer to OKF bundle
5	Model sync	✅ Shipped	P0	OKF config auto-updates from Hermes config.yaml
6	Concept inspector (show command)	✅ Shipped	P0	Inspect any concept with metadata + body
7	CI/CD gating	✅ Shipped	P0	Tests must pass before PyPI publish; skip duplicates
8	Async I/O	🚧 Not started	P2	All file ops are sync today. Needed for high-throughput agents.
9	Git-backed history	🚧 Not started	P1	Bundle is markdown files — git init works manually. No programmatic diff/rollback yet.
10	Web viewer	🚧 Not started	P1	CLI-only today. A small FastAPI/HTML viewer would unlock graph exploration.
11	Plugin system (custom types/validators)	🚧 Not started	P2	OKFValidator is hardcoded. No custom concept types or validator hooks.
12	Multi-agent merge	🚧 Not started	P2	Each agent is isolated. No merge_bundles() or conflict resolution API.
13	Hermes orchestration	🚧 Not started	P3	Single agent per bundle. No multi-agent coordinator.

Legend:

• ✅ Shipped — in hermes-okf v0.5.0
• 🚧 Not started — on the backlog; not planned for 0.5.x

Current focus: v0.5.0 unifies the CLI (install-plugin/uninstall-plugin subcommands) and fixes the critical memory provider integration. Roadmap items 8–13 are for a 1.0 release.

---

License

MIT — see LICENSE.

---

Acknowledgements

Built for the Hermes agent ecosystem and inspired by the Google Cloud Knowledge Catalog team's OKF draft specification. If you use Hermes or OKF, this library is designed to be your memory backbone.

⭐ Star this repo if you're building agent memory systems — let's make Hermes the best agent framework out there.