Pack local or remote codebases into AI-friendly files that LLMs and coding agents can read or search
Rust Anything is installed from its publisher's own source and answers where it runs, so this marketplace is not in the path of a single call. There is no address here to send one to, and a panel that pretended otherwise would be showing you an answer we made up. Install it and call it from your own client — the Installation tab has the entry for each one.
What it does
Rewrite an app in Rust one piece at a time, checked against the original's recorded behavior.
Quickstart
# 1 — add it to Claude Code
claude mcp add --scope user rust-anything-mcp '--' npx -y rust-anything mcp
# 2 — ask your agent for something these tools do
# analyze, bridge, refresh and 13 more
Rust Anything costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.
Collected from a public index. Nobody has claimed this account, so nothing here was written by its author — claim it if it is yours.
Where are you running it?
Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts Rust Anything as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. Its scan result is on the Security tab — read the surface before you approve it.
This is a public server: you run it yourself and this marketplace is not in the path. You need Node.js — the current LTS (24) is the safe choice; npx comes with it. Claude Code registers it in one command; --scope user makes it available in every project. The '--' is in quotes so that PowerShell passes it on.
claude mcp add --scope user rust-anything-mcp '--' npx -y rust-anything mcpReconnect, or start a new session, and the tools appear in the model’s tool list.
One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.
16 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.
Tool surface
What the model actually sees. Descriptions are diffed on every release — see version history.
Read from its source code. A server that builds its tool list when it starts can list more tools than this.
Map a repository (TypeScript/JavaScript, Python, C, Go, Ruby, Java or Kotlin): units, port order (waves), risks with file:line, bridge per unit, dependency-to-crate fit, and start_here: the best first ports.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| root | string | yes | Repository root. |
| lang | string | no | Source language; detected from the repository when omitted. |
| whole_app | boolean | no | Plan a whole-app port: entry points and shape (cli, server, library), every dependency imported or declared with its crate and fit, pieces in port order (units that import each other together), and nu |
Scaffold a Rust crate for one unit: napi-rs or wasm-bindgen bindings in src/lib.rs, stubs to port in src/port.rs, UTF-16 helpers when the unit uses string offsets. Never overwrites an existing folder.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| root | string | yes | — |
| unit | string | yes | Unit id from analyze. |
| kind | string | no | Defaults to the plan's bridge. |
| out | string | no | Crate folder; defaults to <root>/rust/<name>. |
| name | string | no | Crate name; defaults to <unit>-rs. |
| force | boolean | no | Bridge a unit the plan marks blocked or keep. |
| keep | array | no | Exports to keep in the source language: name or name: reason. |
| batch | array | no | Batch classes, function:param: the function runs over many param values kept in Rust, on all cores, in one call. A <Function>Index struct in port.rs replaces the default per-item loop. |
| as_json | array | no | Type parameters the app only fills with data (T in paginate<T>(items: T[])): bridged as JSON values instead of keeping the export in TS. |
| allocator | string | no | Global allocator for crates with batches. mimalloc (default) is C and leaves wasm alone; rusty_alloc is pure Rust (no C compiler) and also serves wasm. |
Regenerate a crate's bindings (lib.rs, index.cjs, js.rs) from the signatures in src/port.rs, after changing them, adding any new batch classes. Never touches port.rs.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| crate | yes | — | |
| batch | array | no | Batch classes to add, function:param: the function runs over many param values kept in Rust, on all cores, in one call. A <Function>Index struct in port.rs replaces the default per-item loop. |
Run the original functions on generated edge-case inputs (non-ASCII, emoji, -0, NaN, ...) plus optional extra inputs, and save the results as cases.jsonl: the behavior the port must match.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| crate | yes | — | |
| count | integer | no | Generated cases per function (default 200). |
| inputs | string | no | JSONL file of extra cases: {"fn": name, "args": [...]}. |
| module | string | no | JS only: import the original from this JS module instead of the TS source. |
| resolve_from | string | no | Resolve npm packages from this folder's node_modules. |
| float_tolerance | number | no | Numeric code: floats match when they differ by at most this times the larger magnitude (absolute below 1), e.g. 1e-9. Saved with the crate for diff and gate; 0 removes it. |
| node_args | array | no | Extra arguments for node, e.g. ["--import=tsx"]. |
| test_command | array | no | TS/JS or Python: also record every call the original's own test suite makes to the unit's functions, e.g. ["npx", "jest"] or ["pytest"]. Runs in the original's folder. |
| profile | string | no | Record on this machine (default) or on a connected Android device (JVM crates). |
| device | string | no | Android device serial; picked automatically when only one is connected. |
Build the crate: the napi-rs .node, wasm-bindgen output, Python extension or shared library. With profile: "android", build the JNI library for Android ABIs into jniLibs/ (needs the NDK and the Rust Android targets; returned as {ndk, api, libraries} instead of {artifact}).
| Parameter | Type | Required | Notes |
|---|---|---|---|
| crate | yes | — | |
| profile | string | no | Build for this machine (default) or for Android. |
| abi | array | no | Android ABIs; defaults to arm64-v8a, armeabi-v7a and x86_64. |
| api | integer | no | Android API level (at least 26; default 26). |
| ndk | string | no | Android NDK root; found from ANDROID_NDK_HOME or the SDK when omitted. |
| jni_libs | string | no | Where to install the Android libraries; defaults to <crate>/jniLibs. |
Replay cases.jsonl against the built Rust module and list mismatches. With profile: "android", replay on the device against the .so built for its ABI.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| crate | yes | — | |
| profile | string | no | Diff on this machine (default) or on a connected Android device (JVM crates). |
| device | string | no | Android device serial; picked automatically when only one is connected. |
| jni_libs | string | no | Android libraries folder to diff; defaults to <crate>/jniLibs. |
| ignore_key_order | boolean | no | — |
| module | string | no | Test this JS module instead of the crate's index.cjs. |
| node_args | array | no | Extra arguments for node, e.g. ["--import=tsx"]. |
| float_tolerance | number | no | Floats match within this tolerance for this run (the one saved by record otherwise). |
Run a real workload (a command: node or python3 script, go test -bench, java, ruby) under the language's profiler and rank the repository's hot functions by inclusive time, with their callers, call counts and advice: what to port together, and which loops to batch (bridge with batch).…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| root | string | yes | — |
| command | array | yes | The workload, run in root: ["node", "bench.mjs"]. |
| lang | string | no | Detected from the repository when omitted. |
Time the recorded cases on the original and on the built Rust module, bridge included; speedup below 1 means the port is slower as its callers call it. Returns the rows and a verdict with advice: ship, keep the slower functions, or port a bigger unit.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| crate | yes | — | |
| resolve_from | string | no | — |
| min_ms | integer | no | Minimum measured time per function and side, in milliseconds (default 200). |
| node_args | array | no | Extra arguments for node, e.g. ["--import=tsx"]. |
Generate a drop-in shadow (a TS or Python module, Go package, Ruby file, Java classes or C file) that runs the original and the Rust port side by side and reports mismatches; RUST_ANYTHING_MODE=old|shadow|new switches at runtime.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| crate | yes | — | |
| out | string | no | — |
Port map of a repository: each unit of the original with the crate that ports it, its functions in Rust and kept in the original, its cases and last diff, the whole app's glue (Rust only, or still handing cases to the original), the units not ported yet in the order analyze suggests, and every…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| repo | string | yes | Repository root. |
| crates | array | no | Folders with crates kept outside the repository. |
Definition of done: fmt, clippy -D warnings, cargo test, unsafe with SAFETY comments only, no unimplemented functions, build, API present, recorded behavior matches.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| crate | yes | — | |
| node_args | array | no | Extra arguments for node, e.g. ["--import=tsx"]. |
Record the whole CLI app's behavior: run the original on its own documented examples, an inputs file and semantic traps (non-ASCII and emoji, negative division, overflow, float extremes, map order), twice each. What differs between runs is reported by kind in noise: values (now [..] wildcards),…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| repo | string | yes | Repository root. |
| run | array | no | Command that runs the original, paths relative to repo: ["python3", "cli.py"]. Optional after the first record. |
| inputs | string | no | JSONL of extra cases: {"name", "args": [...], "stdin", "env": {...}, "files": {path: text}}. |
| normalize | array | no | Normalizers: timestamps, uuids, key-order, or name=regex. |
| name | string | no | Name the app is invoked by in its docs. |
| timeout | integer | no | Seconds before a run is killed (default 30). |
| traps | boolean | no | Add semantic-trap cases (default true). |
| http | boolean | no | The app is a server: run starts it (it takes its port from PORT or {port}); cases are requests from curl lines in its docs, inputs ({"name", "method", "path", "headers", "body"}) and traps. |
| port | integer | no | A port the server always listens on (implies http). |
| effects | array | no | Record what leaves the process: "http" (a local sink set as HTTP_PROXY, http_proxy and RUST_ANYTHING_HTTP_SINK) and "smtp" (a local mail server, SMTP_HOST and SMTP_PORT). Each case stores what reached |
| effect_env | array | no | Point the app's own settings at a sink: VAR=http-url, VAR=smtp-host, VAR=smtp-port or VAR=smtp-url. |
| db_reset | string | no | Shell command that resets the database before each case (each start for servers); cases then run one at a time. |
| db_dump | string | no | Shell command that prints the database's state (e.g. sqlite3 app.db .dump), run after each case or request and recorded with the effects. |
| args | array | no | Servers: arguments every start gets after its command, the original's and every build's (["db.json"]). Implies http; saved. |
| fresh | array | no | Servers: repository files or folders the server reads or writes, snapshotted and copied into a fresh working directory before every start. Implies http; saved. |
| pace | integer | no | Servers: milliseconds to wait after each response, for work the server does after answering (a file watcher). Implies http; saved and used by replay; 0 removes it. |
| float_tolerance | number | no | Numeric output: numbers match when they differ by at most this times the larger magnitude (absolute below 1), e.g. 1e-9. Saved with the cases; 0 removes it. |
| float_file | array | no | Written files of raw floats, compared value by value within float_tolerance: PATH=f32le or PATH=f64le. Saved with the cases. |
Replay <repo>/rust/app-cases/ against a build (the original, the front door, the final Rust binary): passed/total, fallback rate when the build reports it, and a report with a diff per failing case (rust/app-cases/replay-report.md).
| Parameter | Type | Required | Notes |
|---|---|---|---|
| repo | string | yes | Repository root. |
| build | string | no | Command line of the build to test; the original when omitted. |
| port | integer | no | Servers: the port the build always listens on; otherwise it gets a free one through PORT and {port}. |
| timeout | integer | no | Seconds before a run is killed (default 30). |
| case | string | no | Replay only this case, by the name the report gives it (a library: only this unit): the fast loop for fixing one mismatch. |
Scaffold <repo>/rust/app, the Rust binary that owns the CLI's entry point: a clap parser for the recorded commands and flags, every command running the original with the same argv, stdin and env (each fallback counted). For a server: an axum app that starts the original on a port of its own and…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| repo | string | yes | Repository root. |
| check | boolean | no | Build and replay (default true). |
| library | boolean | no | Treat the repository as a library even if it has a main: the front door is a Rust crate with the public API, re-exporting each wired unit's port. Found by default when there is no main and no server. |
After a unit passes gate, hand its commands to its port.rs: adds the unit crate to rust/app as a library, writes a glue stub (src/units/<unit>.rs, yours to write: parse args, call <crate>::port, print what the original prints), regenerates the routing with a kill switch…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| repo | string | yes | Repository root. |
| unit | string | yes | Unit crate folder, its name under rust/, or the unit id it was bridged from. |
| commands | array | no | Commands the unit answers; kept from the last wire when omitted. |
| routes | array | no | Servers: routes the unit answers, axum syntax: "GET /users/{id}", "/path" for any method, "*" for everything else; kept from the last wire when omitted. |
| skip_gate | boolean | no | Do not run gate on the unit first. |
| check | boolean | no | Build and replay (default true). |
Make rust/app a pure Rust app. Refuses (and lists every blocker) unless replay passes with fallback rate 0 and every unit passes gate. Then, on a copy first: removes the fallback and kill switches, turns each unit into a plain library (no cdylib, bindings or generated wrappers), adds a trycmd test…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| repo | string | yes | Repository root. |
- Every tool, no call limit
- No card, and mcprush charges nothing for it
- Source published under MIT
- Runs on your machine — nothing of it reaches our gateway
- Nothing to cap, because mcprush bills nothing
What counts against your monthly calls
| Tool | Unit | Calls used | Out of the allowance |
|---|
Nothing here is billable: mcprush charges nothing to install Rust Anything or to call it, at any volume.
Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.
| Term | Level | What it prices | Points |
|---|---|---|---|
| capability-exposure | high | capability blast radius (high) — client exposure if the model is manipulated | −6 |
| verification-discount | source | publisher verification (provenance) — cryptographic build provenance ties the artifact to its source | −0 |
| coverage-honesty | source | inspection depth (source) — how much of the target the scan could see | −0 |
What the scan could actually read
A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.
No tools were enumerated, so prompt-injection, capability and toxic-flow analysis had no tool surface to inspect. To grade a package’s real runtime tools, scan the running server: --command "npx -y <package>".
Capability — what it could do if the model were manipulated
Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. high is the level these add up to.
| Tool | Capability tags | Why the tag was assigned |
|---|---|---|
| No per-tool rows for this release — see the note below. | ||
Toxic-flow graph
The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.
The public result for this release does not print the flow graph, so there is nothing to show here. That is not the same as "no paths were found": what the scan did read is above, under coverage.
Supply chain and provenance
This is the first scan of this surface here, so there is nothing yet to compare it against.
Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.
Release history
Always latest — the install command names no version, so it installs the newest release when it runs; the current release on file is 2.1.0. Pin a release below and the command asks for exactly that one.
No release note was published with this version.
One review per account, from a signed-in account that does not publish this listing. Nothing else is asked — you can say what you think before you install it. Publishers can reply once per review, and a review can be edited or deleted by whoever wrote it.
Any signed-in account that does not publish Rust Anything can review it, once — before installing it or after. A review from an account that has installed it is marked as one.
Nobody has reviewed this listing. The rating on the card is the mean of the reviews written here and nothing else, so there is no rating until somebody writes the first — which takes a signed-in account that does not publish it, and nothing else.