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.
npx agentmods add instructions/djbelieny/nova/claude-mdgit clone --depth 1 https://github.com/djbelieny/novaWrote 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/instructions/djbelieny/nova/claude-md)<a href="https://agentmods.dev/instructions/djbelieny/nova/claude-md"><img src="https://agentmods.dev/badge/instructions/djbelieny/nova/claude-md.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 | $0.04199 | $0.04199 |
| Opus 5 | $0.02099 | $0.02099 |
| Sonnet 5 | $0.00840 | $0.00840 |
| Haiku 4.5 | $0.00420 | $0.00420 |
Grade C, and why
nova CLAUDE.md scanned grade C with 2 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 2d 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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
<!-- rtk-instructions v2 --> Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
rtk curl <url> # Compact HTTP responses (70%) How it starts
The opening of the file, as written. The whole thing — 384 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Nova — Architecture & Conventions
Claude Code reads this file on every conversation. It describes how Nova actually works so you can navigate the codebase, make changes, and debug without re-exploring each time.
Nova is a multi-agent AI orchestration platform. 24 specialist agents, task decomposition with dependency resolution, human-in-the-loop approval gates, pattern caching, multi-channel messaging (Telegram/WhatsApp/Slack), and 12 MCP integrations.
For setup instructions, see SETUP.md. For production deployment, see DEPLOY.md. For deep technical reference, see docs/ARCHITECTURE.md.
Message Flow
User ──▶ Channel Adapter ──▶ relay.ts ──▶ orchestrator.ts
(Telegram/ │
WhatsApp/ ┌───────┴────────┐
Slack) │ Classify │
│ (3-tier) │
└───────┬────────┘
┌───────────────┼───────────────┐
simple routed complex
│ │ │
callClaude() agent-router.ts planner.ts
(+ memory (specialist (decompose →
context) prompt) parallel exec)
│ │ │
└───────────────┼───────────────┘
│
Two-Phase Execution
prepare ──▶ approval ──▶ execute
│
Response ──▶ User
Key Modules
| File | Role |
|---|---|
src/relay.ts |
Entry point. Channel setup, AI provider registration, message coordination, workspace management |
src/orchestrator.ts |
3-tier classification, approval gates, revision sessions, pattern cache lookup |
src/planner.ts |
Task decomposition (LLM + deterministic plans), parallel execution, artifact aggregation |
src/agent-router.ts |
Agent catalog loading, tool/skill mapping, prompt construction per specialist |
src/memory.ts |
Intent tag parsing, context injection (facts, goals, tasks, history, schedule) |
src/db.ts |
Split SQLite architecture (shared.db + per-user DBs), schema, migrations |
src/ai-provider.ts |
Provider interface & registry (Claude, Gemini, Codex, Groq) |
src/ai-router.ts |
Smart routing: force override → user pref → hint → MCP dep → rate limit fallback |
src/patterns.ts |
Pattern caching: signature normalization, keyword overlap matching, success tracking |
src/channels/ |
Platform adapters: telegram.ts, whatsapp.ts, slack.ts, index.ts (registry + types) |
src/providers/ |
AI CLI wrappers: claude.ts, gemini.ts, codex.ts, groq.ts |
src/integrations.ts |
Per-user MCP config generation, credential management |
src/embeddings.ts |
Local embeddings via all-MiniLM-L6-v2, semantic search |
src/scheduler.ts |
Scheduled task execution (recurring, one-time, conditional) |
services/ |
Proactive services (smart-checkin, morning-briefing, ai-news, social-post, lead-suggester, etc.) |
src/dashboard.ts |
Admin dashboard API |
src/voice-server.ts |
Voice call handling |
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.
- 2d ago First seen · 384 lines · 4,199 tokens per session scan C 38bf38a69b26
nova CLAUDE.md is an instructions file published in the GitHub repository djbelieny/nova (5 stars, last pushed 1mo ago), licensed MIT. It adds 4,199 tokens to every session, about $0.0210 per session on Opus 5. A static security scan graded it C with 2 findings (hidden instructions, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other instructions, from other repositories
commonly CLAUDE.md
Claude Code instructions for Team-Commonly/commonly, covering claude.md / agents.md, 🧠 product vision & architecture philosophy, what commonly is, the architecture model and key concepts.
PuPu AGENTS.md
AGENTS.md instructions for haoxiang-xu/PuPu, covering gitnexus — code intelligence, always do, never do, resources and cli.
PuPu CLAUDE.md
Claude Code instructions for haoxiang-xu/PuPu, covering gitnexus — code intelligence, always do, never do, resources and cli.
comis CLAUDE.md
Claude Code instructions for comisai/comis, covering claude.md, generic runtime check — before every code change, tests-first, docs-current and root-cause before patching.
anomalia AGENTS.md
AGENTS.md instructions for anomaliaso/anomalia, covering anomalia — social media ai autopilot, cli, install (standalone binary, no runtime needed), quick commands and architecture (this repo — the server side of the cli).
box0 CLAUDE.md
Instructions for risingwavelabs/box0, covering claude.md, project, conventions, architecture and resource model.