hermes-cashew

A Hermes Agent memory provider plugin that stores conversation context in a local Cashew thought graph with semantic search and automatic context recall. Get from zero to a working install in under five minutes.

v0.9.0 auto-generates cashew.json with defaults on first load, enables LLM-powered extraction by default (no manual auxiliary.memory setup), and adds forest-level insight extraction via on_pre_compress. v0.8.0 re-enabled the sleep cycle with a ground-up refactored implementation — vectorized cross-linking, batched DB writes, ~4s at 7K nodes.

Prerequisites

• Hermes Agent installed
• cashew-brain>=1.0.0 — installed automatically by hermes plugins install
• sqlite-vec — enables vector similarity search. Installed automatically.

Install

hermes plugins install magnus919/hermes-cashew

This clones the repository to ~/.hermes/plugins/cashew/ and registers the plugin entry point. After install, restart the gateway:

hermes gateway restart

Register with Hermes

After installing, set cashew as the active memory provider:

hermes config set memory.provider cashew
hermes gateway restart

Or use the interactive setup (v0.2.0 now includes cashew in the provider picker):

hermes memory setup

Zero-Config Startup

hermes-cashew works out of the box — all 32 configuration keys have sane defaults. On first agent startup, the plugin auto-generates ~/.hermes/cashew.json with the full default configuration and auto-populates auxiliary.memory in Hermes config.yaml from the main model config, so LLM-powered extraction is active without any manual setup.

Created ~/.hermes/cashew.json only if you want to override specific defaults. The file is never overwritten once it exists:

# Optional: override individual defaults
cat > ~/.hermes/cashew.json << 'EOF'
{
  "recall_k": 10,
  "think_interval": 15,
  "user_domain": "user"
}
EOF

Full Config Reference

Core

Key	Default	Description
cashew_db_path	cashew/brain.db	Path to SQLite DB, relative to hermes_home
embedding_model	thenlper/gte-large	Sentence-transformers model for embeddings (1024-dim)
llm_aux_role	memory	Hermes auxiliary role for LLM-powered extraction; requires auxiliary.memory in config.yaml
auto_extraction	true	Auto-extract knowledge from conversation turns
sync_queue_timeout	30.0	Seconds to wait for sync worker drain on shutdown

Retrieval

Key	Default	Description
recall_k	5	Context fragments returned per query
similarity_threshold	0.3	Minimum similarity for BFS graph walk
walk_depth	2	Graph BFS traversal depth
token_budget	2000	Max tokens per context injection
prefetch_k	3	Nodes to pre-warm into context on each turn
prefetch_cues	3	Cue phrases to send to LLM for prefetch generation

Domains & Classification

Key	Default	Description
user_domain	user	Domain label for user messages
ai_domain	ai	Domain label for AI messages
default_domain	general	Fallback domain for unclassified content
auto_classify	true	Auto-classify nodes into domains
domain_classifications	["personal", "work", "projects", "learning", "system"]	Available domain labels
domain_separation_enabled	true	Enforce domain boundaries in retrieval

Sleep Cycle

Key	Default	Description
sleep_cycles	true	Enable the refactored sleep cycle (cross-linking, dedup, GC, dreams)
sleep_schedule	"every 12h"	Cron schedule for sleep cycle
sleep_max_nodes	2000	Max nodes per sleep cycle tick
think_cycles	true	Enable periodic insight generation (think cycle)
think_interval	10	Turns between think cycle runs (0 = disable)
think_cycle_nodes	5	Node clusters per think cycle
max_think_iterations	3	Max iterative refinements per think cycle
novelty_threshold	0.82	Minimum novelty score to surface an insight

Garbage Collection

Key	Default	Description
gc_mode	soft	"soft" or "hard" decay
gc_threshold	0.05	Minimum importance score before decay
gc_grace_days	7	Days before a node can be decayed
gc_protect_types	["seed", "core_memory"]	Node types exempt from decay
gc_think_cycle_penalty	1.5	Importance penalty multiplier for think-cycle nodes
decay_pruning	true	Prune low-value nodes over time
pattern_detection	true	Detect recurring patterns in extracted knowledge

Tuning

Key	Default	Description
access_weight	0.2	Weight of access count in importance scoring
temporal_weight	0.1	Weight of recency in importance scoring
clustering_eps	0.35	DBSCAN epsilon for think-cycle clustering
clustering_min_samples	3	Minimum samples per cluster in think cycle

Feature Flags

Experimental features gated behind boolean toggles. All default to false. Enable in cashew.json under the _features key:

{"_features": {"experimental_batch_sync": true}}

Key	Default	Description
experimental_batch_sync	false	Drain up to 8 sync turns per worker iteration instead of one-at-a-time
experimental_parallel_retrieval	false	Use parallel retrieval paths for semantic search

Environment variables override config values: prefix any key with CASHEW_ (e.g. CASHEW_RECALL_K=10).

Verify the Install

hermes gateway restart   # ensure gateway picks up the new plugin
hermes memory status

Expected output shows Provider: cashew with Plugin: installed and Status: available.

How It Works

hermes-cashew provides two LLM-accessible tools:

• cashew_query — searches the local thought graph for context relevant to the current conversation. Uses sqlite-vec for semantic search.
• cashew_extract — explicitly persists a conversation turn into the graph. The agent can call this when it judges a turn contains worth-remembering knowledge.

Both tools are registered automatically when Hermes loads the plugin. On each session start, prefetch() retrieves relevant context from the graph and injects it into the system prompt.

Privacy Controls (Optional)

Nodes in the thought graph can carry tags. The cashew_query tool accepts an exclude_tags parameter to filter out nodes with specific tags from results:

{"query": "prior decisions", "exclude_tags": ["vault:private"]}

This works in both the vector search and keyword fallback paths. Common use cases:

• Privacy: Tag sensitive nodes with vault:private to exclude them from group or shared contexts
• Domain isolation: Exclude nodes from specific domains during broad queries
• Declassification: Remove exclusion to reveal previously private nodes

LLM Integration

hermes-cashew enables LLM-powered extraction by default — llm_aux_role is set to "memory", and the plugin auto-populates auxiliary.memory in Hermes config.yaml from the main model config on first load.

No manual configuration is needed. To verify LLM extraction is active:

grep "using" ~/.hermes/logs/agent.log | grep "llm_aux_role"
# Expected: llm_aux_role='memory': using <provider> <model> via <base_url>

To disable LLM extraction (heuristic-only mode), set llm_aux_role to null in cashew.json:

{"llm_aux_role": null}

Or remove the section entirely — the default will regenerate it on next start.

What the LLM enables upstream

• LLM extraction — structured knowledge extraction with typed nodes, confidence scores, tags, and domain assignment
• Think cycles — cross-domain synthesis, generates insight nodes from clusters of related knowledge. Runs every think_interval sync turns (default 10). Set think_interval to 0 to disable.
• Sleep synthesis — Graph consolidation pipeline: cross-linking, dedup, garbage collection, permanence evaluation, core memory promotion, and LLM-powered dream generation. Runs as a Hermes cron job on a configurable schedule (default: every 12 hours), not at session boundaries. The cron script reads cashew.json at runtime and operates without an LLM — if LLM-powered dream synthesis is desired, it requires additional configuration (see Sleep Cycle Cron Scheduling below). Processes up to sleep_max_nodes per cycle (default 2,000).
• Pre-compress insight extraction — Before context compression discards old messages, extracts conversation-arc patterns (topic shifts, framing changes, implicit decisions) using a dedicated LLM prompt. Creates insight/observation nodes in the graph. Requires llm_aux_role configuration. Silent-degrades without LLM.

Without llm_aux_role, the plugin uses heuristic-only extraction — no API calls, no LLM cost, zero-config.

Design note: The auxiliary.memory convention is provider-agnostic. Any memory provider plugin can declare llm_aux_role and reference the same auxiliary.memory section, making this a standard pattern across the Hermes plugin ecosystem.

Sleep Cycle Cron Scheduling

hermes-cashew runs its graph consolidation pipeline (cross-linking, dedup, garbage collection, permanence evaluation, core memory promotion) as a Hermes no_agent cron job, not at session boundaries. This means /new returns instantly — no synchronous sleep cycle work blocks the start of a new session.

When the cron job is registered

The cron job is created during plugin initialization (initialize()) only when all of the following are true:

Condition	Config Key	Default	Behavior if false
Sleep cycles enabled	sleep_cycles	true	Cron not registered
Schedule non-empty	sleep_schedule	\"every 12h\"	Cron not registered; set to \"\" to disable
Provider init succeeds	—	—	Exception caught, _config set to None, cron never reached
Hermes cron module available	—	—	ImportError caught, WARNING logged
create_job() succeeds	—	—	Exception caught, WARNING logged
No job already registered for this provider instance	—	—	No-op dedup guard

The cron job is removed on plugin shutdown (shutdown()). A dedup helper scans for existing cashew-sleep-cycle jobs by name on each registration to prevent N jobs accumulating across N restarts.

When the cron job runs

On the configured schedule (default every 12h), the Hermes scheduler executes $HERMES_HOME/scripts/cashew-sleep-cycle.py with no LLM — it is a no_agent script, meaning zero LLM overhead per tick. The script reads cashew.json at runtime to discover its database path and sleep_max_nodes setting.

What happens during a cron tick

1. Reads cashew.json to get cashew_db_path and sleep_max_nodes
2. Selects up to sleep_max_nodes (default 2,000) oldest-unprocessed nodes
3. Computes pairwise cosine similarity (vectorized numpy)
4. Creates cross-links between similar node pairs (threshold: 0.78)
5. Deduplicates near-identical nodes (threshold: 0.82) via BFS clustering
6. Runs garbage collection on low-fitness isolated nodes
7. Promotes frequently-accessed nodes to permanent / core memory status
8. Prints a JSON summary (captured by the cron scheduler's output log)

No LLM-powered dream generation occurs in cron mode — the script passes model_fn=None. Cross-linking, dedup, and GC are the 80% benefit without the API key dependency in a subprocess.

Config reference

Key	Default	Description
sleep_schedule	\"every 12h\"	Cron expression or interval string. Set to \"\" to disable cron-based scheduling entirely. Examples: \"every 30m\", \"0 */2 * * *\", \"0 3 * * *\" (daily at 3am).
sleep_max_nodes	2000	Maximum number of nodes to cross-link in a single sleep cycle. Higher values converge faster but take longer per tick.

Semantic Search

sqlite-vec enables vector similarity search and is installed automatically as a standard dependency. If your platform doesn't support sqlite-vec's native extension, the plugin degrades gracefully to keyword-based retrieval — still functional, but less precise.

sqlite-vec is a standard dependency and will always be loaded at startup.

Uninstall

hermes plugins remove cashew
hermes config set memory.provider built-in   # revert to built-in memory
rm -rf ~/.hermes/cashew   # optional: remove the local graph data

Troubleshooting

Plugin: NOT installed in hermes memory status

1. cashew-brain not installed in Hermes venv — hermes plugins install does not automatically install Python package dependencies into Hermes's venv. Install it manually:

   ~/.hermes/hermes-agent/venv/bin/python3 -m ensurepip
   ~/.hermes/hermes-agent/venv/bin/python3 -m pip install cashew-brain

2. Stale pycache or entry point not registered — If cashew-brain is installed but the plugin still shows NOT installed:

   cd ~/.hermes/plugins/cashew && \
     ~/.hermes/hermes-agent/venv/bin/python3 -m pip install -e .
   hermes gateway restart

Status: not available

The plugin is available when cashew-brain is importable. Check:

~/.hermes/hermes-agent/venv/bin/python3 -c "from core.context import ContextRetriever; print('ok')"

If this fails, cashew-brain is not installed in the Hermes venv (see above).

Hermes-agent venv has no pip

Hermes-agent creates a minimal venv without pip. Bootstrap it first:

~/.hermes/hermes-agent/venv/bin/python3 -m ensurepip
~/.hermes/hermes-agent/venv/bin/python3 -m pip install <package>

Do not run pip install from outside the venv targeting the hermes python, or the package will land in the wrong environment.

Embedding model download on first use

cashew-brain bundles sentence-transformers. The first retrieval operation may trigger a ~500 MB embedding model download. To avoid this in automated environments:

HF_HUB_OFFLINE=1 TRANSFORMERS_OFFLINE=1 HF_DATASETS_OFFLINE=1 hermes ...

Development

git clone https://github.com/magnus919/hermes-cashew
cd hermes-cashew
pip install -e ".[dev]"   # macOS
python3 -m pip install -e ".[dev]"   # Linux
pytest                      # run the test suite

Tests require no network access and mock the embedding model automatically (HF_HUB_OFFLINE=1 is set by tests/conftest.py).

Architecture Notes

The plugin uses a dual-path loading strategy to support both pip install -e . (development) and hermes plugins install (flat-entry loader):

• pip / test path: Python's namespace package mechanism resolves plugins.memory.cashew to plugins/memory/cashew/__init__.py via sys.path
• flat-entry path: Hermes loads ~/.hermes/plugins/cashew/__init__.py as _hermes_user_memory.cashew. The root __init__.py detects this context and exec's the nested implementation with sys.modules patched so relative imports resolve correctly

License

See LICENSE.