Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/matthewdcage/llm-swarm-routerWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/matthewdcage/llm-swarm-router/honcho_rules)<a href="https://agentmods.dev/rules/matthewdcage/llm-swarm-router/honcho_rules"><img src="https://agentmods.dev/badge/rules/matthewdcage/llm-swarm-router/honcho_rules.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.01412 | $0.01412 |
| Opus 5 | $0.00706 | $0.00706 |
| Sonnet 5 | $0.00282 | $0.00282 |
| Haiku 4.5 | $0.00141 | $0.00141 |
Grade A, and why
honcho_rules scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 158 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Honcho Memory — Capture, Recall, and Consolidate
You have access to the honcho MCP server at http://127.0.0.1:8787.
It stores conversation history and builds a persistent psychological model of the user over time.
State File
Always read this file first — it tells you everything you need to skip redundant MCP calls:
.cursor/hooks/state/honcho-state.json
It contains the active session ID, when context was last loaded, turn counts, and dream thresholds. Write back to it after key operations.
Identity
| Role | Peer ID |
|---|---|
| User (human) | matthewcage |
| You (assistant) | Assistant |
| Workspace | workspace.id (from state file) |
On Every Conversation Start
Note: The
.cursor/hooks/honcho-init.pysessionStart hook runs automatically before your first response. It handles session creation and state file maintenance. The hook injects context viaadditional_context(Cursor) /additionalContext(Claude Code) — you will see its HONCHO SESSION READY block in your initial context if hooks are working.
Step 1 — Read state file (always):
Read .cursor/hooks/state/honcho-state.json. The hook has already updated it, so
session.id is guaranteed to be today's session and session.date will equal today.
Step 2 — Session creation (hook already handled this):
The hook called the Honcho API to create cursor-{today} and add both peers if this is a new day.
You only need to call create_session / add_peers_to_session if the state file shows
session.date != today (API was unreachable when the hook ran — run ./start.sh).
Step 3 — Load user context (your job — hook cannot do this):
if state.memory.lastContextLoadedAt is null OR date(lastContextLoadedAt) != today:
get_peer_context(peer_id: "Assistant", target_peer_id: "matthewcage",
search_query: "current projects, preferences, working style")
→ Update state: memory.lastContextLoadedAt = now()
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 158 lines · 1,412 tokens per session scan A 9af632d6c229
honcho_rules is a cursor rule published in the GitHub repository matthewdcage/llm-swarm-router (25 stars, last pushed 5d ago), licensed MIT. It adds 1,412 tokens to every session, about $0.0071 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other cursor rules, from other repositories
compression-safety
Cursor rule "compression-safety" from yvgude/lean-ctx, covering compression safety — prevent lean-ctx marker corruption, the problem, rules for all agents, never write compressed content to files and file restoration commands.
nauro-adopt
Seeds Nauro's project store from an existing repo. Use after nauro adopt has run locally. On filesystem-capable surfaces, reads docs (README, manifests, ADRs, Memory-Bank) for rationale and inspects code, config, tests, lockfiles, and recent git history for evidence, then surfaces targeted probes that turn evidence…
nauro-context
Writes durable shared context into Nauro's project store so other agents (a later session or a parallel one) can discover and pull it, finds and reads context another agent left, or captures a resumable brief so your own next session in this environment picks up cleanly. Three modes. Author writes a shared brief for…
memforge-auto-recall
Rules that make an AI assistant automatically retrieve and save team memories during conversations.
compendium
Prefer the Compendium MCP gateway to shrink noisy/large context before pasting it into the conversation.
memforge-topology-trigger
A set of rules for creating and querying a visual map of how services and repositories in a product relate to one another. The map is shown in Memforge WebUI, a browser interface for viewing this information.