Moltis is a persistent personal agent server written in Rust that runs on hardware controlled by its user. It provides an AI agent with sandboxed command execution, model-provider connections, memory, voice, scheduling, messaging integrations, browser automation, and MCP tools. Its catalogue add-ons extend the agent’s workflows and available tools.
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/moltis-org/moltis/claude-mdgit clone --depth 1 https://github.com/moltis-org/moltisWrote 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/moltis-org/moltis/claude-md)<a href="https://agentmods.dev/instructions/moltis-org/moltis/claude-md"><img src="https://agentmods.dev/badge/instructions/moltis-org/moltis/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 | $0.06226 | $0.06226 |
| Opus 5 | $0.03113 | $0.03113 |
| Sonnet 5 | $0.01245 | $0.01245 |
| Haiku 4.5 | $0.00623 | $0.00623 |
Grade A, and why
moltis CLAUDE.md 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 6d 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 — 534 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Rust version of openclaw (docs, code). All code must have tests with high coverage. Always check for security.
Cargo Features
Enable new feature flags by default in crates/cli/Cargo.toml (opt-out, not opt-in):
[features]
default = ["foo", ...]
foo = ["moltis-gateway/foo"]
Workspace Dependencies
Add new crates to [workspace.dependencies] in root Cargo.toml, reference with { workspace = true }.
Never add versions directly in crate Cargo.toml. Use latest stable crates.io version.
Config Schema and Validation
When adding/renaming fields in MoltisConfig (crates/config/src/schema.rs), also update
build_schema_map() in crates/config/src/validate.rs. New enum variants for string-typed
fields need updates in check_semantic_warnings().
Rust Style and Idioms
- File size limit: 1,500 lines. CI enforces this via
scripts/check-file-size.sh. Split large files into modules by domain. Existing oversize files are allowlisted for incremental decomposition. - Do not add implementation code to
mod.rsorlib.rs. Keep those files for module wiring, exports, and crate setup, move real logic into dedicated sibling modules. - Use traits for behaviour boundaries. Prefer generics for hot paths,
dyn Traitfor heterogeneous/runtime dispatch. - Derive
Defaultwhen all fields have sensible defaults. - Use concrete types (
struct/enum) overserde_json::Valuewherever shape is known. - Match on types, never strings. Only convert to strings at serialization/display boundaries.
- Prefer
From/Into/TryFrom/TryIntoover manual conversions. Ask before adding manual conversion paths. - DRY cross-crate types: When two crates need the same enum/struct, define it once in the lower-level crate and re-export via
pub type Alias = other_crate::Typefrom the higher-level one. Never duplicate enums across crates or round-trip through strings (parse(&id.to_string())) to convert between mirror types. - Prefer streaming over non-streaming API calls.
- Run independent async work concurrently (
tokio::join!,futures::join_all). - Never use
block_oninside async context. - Forbidden:
Mutex<()>/Arc<Mutex<()>>— mutex must guard actual state. - Use
anyhow::Resultfor app errors,thiserrorfor library errors. Propagate with?. - Never
.unwrap()/.expect()in production. Workspace lints deny these. Use?,ok_or_else,unwrap_or_default,unwrap_or_else(|e| e.into_inner())for locks. - Use
timecrate (workspace dep) for date/time — no manual epoch math or magic constants like86400. - Prefer
chronoonly if already imported in the crate; default totimefor new code. - Prefer crates over subprocesses (
std::process::Command). Use subprocesses only when no mature crate exists. - Prefer guard clauses (early returns) over nested
ifblocks. - Prefer iterators/combinators over manual loops. Use
Cow<'_, str>when allocation is conditional. - Keep public API surfaces small. Use
#[must_use]where return values matter.
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.
- 6d ago First seen · 534 lines · 6,226 tokens per session scan A e661ed65922c
moltis CLAUDE.md is an instructions file published in the GitHub repository moltis-org/moltis (2,845 stars, last pushed 2d ago), licensed MIT. It adds 6,226 tokens to every session, about $0.0311 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 instructions, from other repositories
autoclaw.sh AGENTS.md
Instructions for datopian/autoclaw.sh, covering agent instructions, quick reference, landing the plane (session completion), issue tracking with bd (beads) and why bd?.
nanoclaw CLAUDE.md
Claude Code instructions for nanocoai/nanoclaw, covering ⚠️ stop — read this first if you are claude ⚠️, nanoclaw, quick context, entity model and two-db session split.
openhuman CLAUDE.md
Claude Code instructions for tinyhumansai/openhuman, a project described as: OpenHuman is an open source personal AI for Mac, Windows and Linux — local-first memory, agent orchestration, and deep research.
lobu AGENTS.md
AGENTS.md instructions for lobu-ai/lobu, covering repo map, unrecoverable — never do these, facts you cannot derive, ship a change and how to work.
clawrun CLAUDE.md
Instructions for clawrun-sh/clawrun, covering claude.md — clawrun, what this is, repo structure, package responsibilities and packages/cli (clawrun).
mnemo CLAUDE.md
Instructions for Logos-Flux/mnemo, covering mnemo - contributor guide, project overview, architecture, tech stack and coding standards.