The official Svelte MCP server providing docs and autofixing tools for Svelte development
Stlink 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
Program, run and inspect STM32 targets over ST-LINK; every result verified, unverified or failed.
Quickstart
# 1 — add it to Claude Code
claude mcp add --scope user stlink-mcp '--' uvx stlink-mcp
# 2 — ask your agent for something these tools do
# list_probes, board_state, run_firmware and 13 more
Stlink 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 Stlink 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 uv (docs.astral.sh/uv): uvx 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 stlink-mcp '--' uvx stlink-mcpReconnect, or start a new session, and the tools appear in the model’s tool list.
Nothing is required, but it can take STLINK_MCP_CLT_ROOT (STM32CubeCLT (or STM32CubeIDE) install directory, when it is not in a standard place), STLINK_MCP_BOARDS (TOML file describing boards this project has not characterised; defaults to ~/.config/stlink-mcp/boards.toml if it exists), STLINK_MCP_NO_VENDOR_SEED (Set to anything to run on measured and user-stated data only, with no seeding from ST's device databases), STLINK_MCP_CONSOLE (Serial console per probe, SERIAL=/dev/ttyUSB0, comma-separated for several; for a console that is not the probe's own), STLINK_MCP_CONSOLE_BAUD (Console baud rate), STLINK_MCP_GDB_PORT (First GDB server port; each probe's server takes this port and the next) if you want to set them.
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.
List every attached ST-LINK with its firmware, board name, serial console and whether a GDB server holds it. Use the console mapping rather than the order of /dev/ttyACM*, which does not follow the probes. Listing takes about a second, and often much longer while a GDB server holds a probe, which…
Takes no parameters.
Identify the target, the probe and its family, and say whether the core is running, sleeping, halted or locked up, from program-counter samples read without halting it; given elf_path, where in the source they fall. If the core cannot be reached as it runs, reconnects through a reset, which…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| probe_serial | string | no | — |
| interface | no | — | |
| elf_path | string | no | — |
Reset the target, load an ELF into its RAM, start it, and verify that it
| Parameter | Type | Required | Notes |
|---|---|---|---|
| elf_path | string | yes | — |
| probe_serial | string | no | — |
| console_seconds | number | no | — |
| interface | no | — | |
| keep_probe | boolean | no | — |
Stop the GDB server holding a probe; with no serial, every probe's. Resumes a core run_until left halted. While held, the reading tools work beside the server, but a core that locks up is reset, clearing its fault registers. A program that closes the debug port closes it held or not; board_state…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| probe_serial | string | no | — |
Program an image into flash: a raw .bin at address or at a measured flash base, or an ELF or HEX file at its own addresses, inside the part's flash; on a base from ST's database or a board file, only inside the main-flash window; the result names the base's provenance. Verified only when every…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| image_path | string | yes | — |
| probe_serial | string | no | — |
| address | integer/string | no | — |
| interface | no | — |
Read target memory without halting or resetting it, given one of: address (an integer or a "0x..." string, 4-byte aligned, length_bytes a multiple of 4, at most 1024); symbol, a variable decoded with the types in elf_path; or register, such as "RCC.CR", decoded from ST's SVD file for the…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| address | integer/string | no | — |
| length_bytes | integer | no | — |
| symbol | string | no | — |
| elf_path | string | no | — |
| register | string | no | — |
| svd | string | no | — |
| read_effects_ok | boolean | no | — |
| probe_serial | string | no | — |
| interface | no | — |
Sample values while the program runs, without halting or resetting it: what lists variables (decoded with elf_path's types; "g_state.mode" picks a member), "0x..." word addresses, or registers such as "ADC3.ISR" (svd and read_effects_ok as in read_memory). All are read together, a few…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| what | array | yes | — |
| elf_path | string | no | — |
| seconds | number | no | — |
| until | string | no | — |
| svd | string | no | — |
| read_effects_ok | boolean | no | — |
| probe_serial | string | no | — |
| interface | no | — |
Run the target until it reaches location (a function, its inlined copies included, "file.c:line" or a "0x..." address) or writes watch (a variable, or a "0x..." address watched as 4 bytes), for at most seconds. The core is left halted where it stopped, with the call stack, until…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| elf_path | string | yes | — |
| location | string | no | — |
| watch | string | no | — |
| seconds | number | no | — |
| probe_serial | string | no | — |
| interface | no | — |
Reset the target and check that it restarted: the program counter is compared with the reset vector during a brief halt, which is reported, and the core is then left running from reset. Needs an open debug port: when the running program has closed it, board_state reconnects through a reset, and…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| probe_serial | string | no | — |
| interface | no | — |
Read the core registers, halting the core for the read; with elf_path, also the call stack with each frame's arguments and locals. Says whether the core runs again afterwards. Works while run_until holds the core halted, and leaves it so. Without elf_path, a core it cannot reach as it runs is…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| probe_serial | string | no | — |
| interface | no | — | |
| elf_path | string | no | — |
Read and decode the Cortex-M fault status registers without halting or resetting the target. They stay set until software or a reset clears them, so call this first after a crash; a probe run_firmware still holds resets a core that locks up, which clears them. Says whether a fault is active, what…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| probe_serial | string | no | — |
| interface | no | — |
Read the option bytes as ST's programmer reads and explains them: readout protection first, then each section, such as brown-out level, user configuration, boot addresses and write protection, with every value and its meaning. Read only: no tool here writes option bytes, since some writes cannot…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| probe_serial | string | no | — |
| interface | no | — |
Find the instruction that faulted, from the exception frame. Halts the core to read registers, and never resets it, since a reset loses the fault: refuses when the core cannot be reached without one. Call it while the core sits in a fault handler. Returns the stacked PC and LR, from the stack that…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| probe_serial | string | no | — |
| elf_path | string | no | — |
| interface | no | — |
Disassemble a function, or the code around an address, from the ELF, with each instruction's source file and line and the function an inlined copy comes from. location is a function name or an address. Without probe_serial it reads only the ELF; with it, it also compares the target's code with…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| elf_path | string | yes | — |
| location | string | yes | — |
| probe_serial | string | no | — |
| interface | no | — |
Capture what a target prints on its serial console for up to 60 seconds.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| seconds | number | no | — |
| probe_serial | string | no | — |
Capture what the target writes to an ITM stimulus port, over the SWO pin.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| core_clock_hz | integer | yes | — |
| seconds | number | no | — |
| itm_port | integer | no | — |
| probe_serial | string | no | — |
- 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 Stlink 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.
Tools were statically extracted from the published source (16 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, 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 |
|---|---|---|
| list_probes | no tags | |
| board_state | no tags | |
| run_firmware | no tags | |
| release_probe | no tags | |
| flash_firmware | no tags | |
| read_memory | no tags | |
| sample_values | no tags | |
| run_until | no tags | |
| reset_target | no tags | |
| core_registers | no tags | |
| read_fault_status | no tags | |
| read_option_bytes | no tags | |
| locate_fault | no tags | |
| disassemble | no tags | |
| capture_console | no tags | |
| capture_swo | no tags |
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 0.9.0. 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 Stlink 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.