DOCA Rivermax
Where to start: This skill assumes DOCA is already installed,
AND that the NVIDIA Rivermax SDK is separately installed with a
valid Rivermax license present on the host. The agent's FIRST
action on any Rivermax question is to confirm both — without
Rivermax SDK + license, doca-rmax cannot function regardless
of how clean the DOCA-side code is, and this is the #1 first-app
confusion (DOCA does not bundle Rivermax; it wraps it). Open
TASKS.md if the user wants to do something
(configure / build / modify / run / test / debug); open
CAPABILITIES.md when the question is what
can DOCA Rivermax express on this version + this Rivermax
install. If the user has not installed DOCA yet, route to
doca-setup first; for the Rivermax
SDK + license install itself, route to the public DOCA Rivermax
guide (slug DOCA-Rivermax) via
doca-public-knowledge-map.
If the user is asking "how do I get packets to land on my
Rivermax input stream at all", the answer is layered:
doca-rmax is the Rivermax integration surface,
doca-eth is the queue surface that
carries the packets, and doca-flow
is the steering surface that directs them.
Example questions this skill answers well
The CLASSES of DOCA Rivermax questions this skill is built to
answer, each with one worked example. The agent should treat the
class as the load-bearing piece — the worked example is a
single instance.
- "Can I even use
doca-rmax on this host?" — worked
example: "the public docs mention doca-rmax; is it
usable without doing anything else?". Answered by the
Rivermax-SDK + license precondition rule in
CAPABILITIES.md ## Safety policy
- the env-prep checklist in
TASKS.md ## configure step 1, which
routes the install-side question to the public Rivermax guide
via
doca-public-knowledge-map
and refuses to recommend a fallback to doca-eth alone that
would silently lose the timing properties.
- "How do I set up a SMPTE ST 2110 video receive stream?" —
worked example: "line up an inbound Rivermax stream on a
representor of a BlueField port for first-run testing".
Answered by the per-stream object lifecycle in
TASKS.md ## configure +
CAPABILITIES.md ## Capabilities and modes
input stream capability table.
- "Which
doca_rmax_get_*_supported query do I have to call before
picking a PTP clock or hardware packet-placement order?" — worked
example: "can this device + this Rivermax install use
ST 2110-20 sequence-number placement?". Answered by the
capability-query rule in
CAPABILITIES.md ## Capabilities and modes
- step 3 in
TASKS.md ## configure.
- "My stream is set up but no packets arrive — why?" —
worked example: "the Rivermax stream object started cleanly
but no recv events fire". Answered by the precondition matrix
in
CAPABILITIES.md ## Safety policy
- the env-prep checklist in
TASKS.md ## configure step 1, which
routes the steering side to
doca-flow, the queue side to
doca-eth, and the license side
back to the Rivermax-side precondition.
- "Is this Rivermax integration capability available on my
device + my installed DOCA + my Rivermax SDK?" — worked
example: "does this device + this Rivermax version advertise
the PTP clock or placement mode I need". Answered by the capability-query
rule in
CAPABILITIES.md ## Capabilities and modes
- the version-and-device overlay in
CAPABILITIES.md ## Version compatibility,
which adds the Rivermax-side version is a second axis rule
on top of the canonical DOCA version-handling chain in
doca-version.
- "What does this
DOCA_ERROR_* from a Rivermax call mean and
which layer caused it?" — worked example:
"DOCA_ERROR_NOT_SUPPORTED from doca_rmax_init()".
Answered by the Rivermax overlay on the
cross-library taxonomy in
CAPABILITIES.md ## Error taxonomy
- the layered ladder in
TASKS.md ## debug that escalates to
doca-debug, and which preserves
the installed header's call-specific mapping: init-time
_NOT_SUPPORTED routes first to Rivermax SDK / license checks;
later errors are interpreted from the exact failing call rather
than generalized into a license diagnosis.
Audience
This skill serves external developers building applications
that consume the DOCA Rivermax integration — i.e., users whose
code calls doca_rmax_* (directly in C/C++, or through
FFI/bindings from another language) to drive timing-precise
media-over-IP streams (SMPTE ST 2110, real-time market data,
high-throughput scientific instrument streams) on top of a
separately-installed NVIDIA Rivermax SDK on a BlueField or
ConnectX host. It is not for NVIDIA developers contributing to
DOCA Rivermax itself, and it is not for users who want
best-effort packet I/O — for that, route to
doca-eth directly.
Language scope. DOCA Rivermax ships as a C library with
pkg-config module name doca-rmax. The shipped samples
live under /opt/mellanox/doca/samples/doca_rmax/ and are
written in C. C and C++ consumers are the canonical case; the
worked examples in TASKS.md assume that path. Other-language
consumers (Rust, Go, Python, …) consume the same *.so through
FFI or language-specific bindings; the skill's contribution in
that case is to keep the precondition rule (Rivermax SDK +
license), per-stream lifecycle, capability-discovery,
permission, scheduling-discipline, and error-taxonomy guidance
language-neutral, and to route the agent to the public C ABI as
the authoritative surface that any wrapper will eventually call.
When to load this skill
Load this skill when the user is doing hands-on DOCA Rivermax
work, in any language. Concretely:
- Confirming that the NVIDIA Rivermax SDK and a valid Rivermax
license are present on the host before any DOCA-side code is
written — this is a mandatory precondition, not an error
path; without it the answer is "
doca-rmax cannot be
used; pick a different library", not "let's try and see what
fails".
- Initializing the global DOCA Rivermax engine with
doca_rmax_init() / doca_rmax_release(), then creating a
doca_rmax_in_stream (receive) context with
doca_rmax_in_stream_create() on a doca_dev opened against a
physical port, a representor, or an SF, and converting it via
doca_rmax_in_stream_as_ctx() before doca_ctx_start(). The
public DOCA Rivermax API is receive-only — there is no
transmit/output stream object.
- Configuring a Rivermax input stream for SMPTE ST 2110 audio +
video over IP, real-time market data feeds, or a scientific
instrument stream — with awareness that PTP-clock and
hardware packet-placement-order support are device- and
Rivermax-conditional and gate on the matching
doca_rmax_get_*_supported query.
- Reading or setting Rivermax stream properties via
doca_rmax_in_stream_set_* and querying device + Rivermax
capability via the doca_rmax_get_*_supported family before
assuming PTP-clock or hardware packet-placement-order support.
- Pairing the Rivermax stream with the
doca-eth queue surface
that carries the packets and the doca-flow rules that
steer them — Rivermax does not program steering itself.
- Designing the real-time scheduling discipline for the
streaming threads (the underlying Rivermax stack expects
real-time priority for sub-microsecond jitter; the canonical
scheduling guidance lives in the Rivermax SDK docs reachable
through
doca-public-knowledge-map).
- Debugging a
DOCA_ERROR_* returned from a Rivermax call
(lifecycle vs. license / permission vs. capability vs.
driver-below) where the cause may live in the DOCA-side
wrapper, the underlying Rivermax stack, or the licensing
layer.
- Designing or extending non-C bindings (Rust, Go, Python, …)
that wrap the DOCA Rivermax C ABI — for the precondition
rule, lifecycle, capability, permission, and scheduling rules
the wrapper must honor.
Do not load this skill for general DOCA orientation, for
installing DOCA itself, for installing the Rivermax SDK or
managing the Rivermax license file (those live in the public
Rivermax SDK guide reachable through
doca-public-knowledge-map),
for best-effort packet I/O without timing requirements (use
doca-eth directly), or for pure
host-side data processing without networking. For DOCA
documentation orientation, use
doca-public-knowledge-map.
What this skill provides
This is a thin loader. The body keeps only the orientation
needed to pick the right next file. The substantive
Rivermax-specific material lives in two companion files:
CAPABILITIES.md — what DOCA Rivermax can express on this
version and this Rivermax SDK install: the
Rivermax-as-hard-dependency rule, the receive-only
doca_rmax_in_stream object model, the capability-query surface
(doca_rmax_get_*_supported), the Rivermax error taxonomy
(mapped onto the cross-library DOCA_ERROR_* set, with
Rivermax-specific causes called out per row), the
observability surface (per-stream progress engine events,
capability snapshots, Rivermax-side license + driver state),
and the safety policy that gates the Rivermax-SDK-present /
license-present / device-access / scheduling-discipline
preconditions.
TASKS.md — step-by-step workflows for the six in-scope
Rivermax verbs: configure, build, modify, run,
test, debug. Plus a Deferred task verbs block that
points out-of-scope questions (installing Rivermax,
managing the license, programming steering, programming the
underlying queue) at the right next skill, and a Command appendix of the recurring commands the agent reaches for.
The skill assumes a host or BlueField where DOCA is already
installed at the standard location, the NVIDIA Rivermax SDK is
installed at its expected location with a valid license, and
the user has the privileges their public install profile
expects (typically sudo or mlnx-group membership to open a
doca_dev against a port). It does not cover installing DOCA —
that path goes through
doca-setup. It does not cover
installing Rivermax or its license — that path goes through the
public Rivermax SDK guide reachable through
doca-public-knowledge-map.
What this skill deliberately does not ship
This skill is agent guidance, not a samples or templates
bundle. To keep the boundary clean, it deliberately does not
contain — and pull requests should not add:
- Pre-written DOCA Rivermax application source code, in any
language. The verified Rivermax source code is the shipped
C samples at
/opt/mellanox/doca/samples/doca_rmax/<name>/.
The agent's job is to route the user to those files and
prescribe a minimum-diff modification on them via the
universal modify-a-sample workflow in
doca-programming-guide,
layered with the Rivermax-specific overrides in
TASKS.md ## modify.
- Standalone build manifests (
meson.build,
CMakeLists.txt, Cargo.toml, …) parked inside the skill.
The agent constructs the build manifest in the user's
project directory against the user's installed DOCA, where
pkg-config --modversion doca-rmax is the source of
truth.
- Quoted Rivermax symbol names beyond the public family
pattern (
doca_rmax_get_*_supported for capability discovery,
receive-only doca_rmax_in_stream session objects driven by the
standard DOCA Core context lifecycle). Exact Rivermax
symbol names are install-bound and Rivermax-SDK-version-
bound; the agent should read them from the installed headers
at $(pkg-config --variable=includedir doca-common) and from the
public DOCA Rivermax guide rather than rely on agent memory.
- A
samples/, bindings/, or reference/ subtree of
any kind. A mock or incomplete artifact in this skill's
tree, even one labeled "reference", is misleading: users
will read it as buildable.
Loading order
- Read this
SKILL.md first to confirm the user's question is
in scope and that the Rivermax-SDK + license
precondition has been considered.
- For the Rivermax capability matrix, the receive-only input
stream model, capability-query rules, error taxonomy,
observability, and safety policy, see
CAPABILITIES.md.
- For step-by-step workflows — configure, build, modify,
run, test, debug — see TASKS.md.
Both companion files cross-link to each other,
doca-version for the canonical
DOCA version-handling rules (Rivermax adds a Rivermax-SDK-
version axis on top),
doca-eth for the queue surface
that carries the packets,
doca-flow for the steering side
that decides which packets land on which queue, and
doca-public-knowledge-map
whenever the right answer is "look it up in the public
Rivermax SDK guide or the installed package layout" rather
than "Rivermax-integration-specific guidance".
Related skills
doca-public-knowledge-map —
the routing table for every public DOCA documentation source
and the on-disk layout of an installed DOCA package. The
Rivermax URL slug is DOCA-Rivermax; the public Rivermax SDK
guide (separate product) is reachable from the same routing
table when the user asks how to install Rivermax or its
license.
doca-setup — env preparation,
install verification, port-state checks (devlink dev show,
ip link), permission and group-membership requirements for
opening a doca_dev. This skill assumes its preconditions
are satisfied; the Rivermax-SDK + license preconditions are
layered on top.
doca-version — canonical
DOCA version-handling rules. This skill's ## Version compatibility cross-links the four-way match rule and adds
the Rivermax-specific overlay (Rivermax SDK version is a
second axis; the capability set on this host is the
intersection of DOCA-side cap-query results and
Rivermax-side capabilities).
doca-structured-tools-contract —
the bundle's structured-tools precedence rule (detect /
prefer / fall back / report). The Command appendix in
TASKS.md honors this contract.
doca-programming-guide —
general DOCA programming patterns shared by every library:
the canonical pkg-config + meson build pattern, the
universal modify-a-shipped-sample first-app workflow, the
universal lifecycle, the cross-library DOCA_ERROR_*
taxonomy, and the program-side debug order. This skill
layers Rivermax specifics on top.
doca-eth — the queue surface
that carries the packets a Rivermax stream produces or
consumes. DOCA Rivermax does not program the underlying
queue itself; it integrates with the queue programmed via
doca-eth. The two skills' lifecycles are independent.
doca-flow — the steering
surface that decides which packets land on which queue.
DOCA Rivermax does not program steering itself; an empty
Rivermax input stream almost always means a missing or wrong
Flow rule (or a missing Rivermax license), not a Rivermax
bug.
doca-debug — the
cross-cutting debug ladder (install / version / build /
link / runtime / program / driver). Rivermax-specific debug
(license precondition gaps, stream-type / packet-rate
capability mismatches, scheduling-discipline jitter
symptoms) overlays on top of that ladder.
1---2name: doca-rmax3description: Use this skill when the user is doing hands-on DOCA Rivermax work on a BlueField DPU or ConnectX host — standing up `doca_rmax_in_stream` (receive) sessions for timing-precise media-over-IP (SMPTE ST 2110 video/audio, market data, scientific feeds), confirming the Rivermax SDK + license precondition before any DOCA-side code, running `doca_rmax_get_*_supported` capability queries, pairing with `doca-eth` queues and `doca-flow` steering, or debugging `DOCA_ERROR_*` from a Rivermax call. Trigger even when the user does not explicitly mention "DOCA Rivermax" or "rmax" — implicit phrasings include "ST 2110 receive isn't getting frames", "sub-microsecond jitter on BlueField", "NOT_SUPPORTED from doca_rmax_init", "no recv events after stream start", or "license check failing on a media receiver". Refuse and route elsewhere for installing the Rivermax SDK or its license, programming the underlying queue (`doca-eth`), steering rules (`doca-flow`), or best-effort packet I/O — those belong to other skills.4license: Apache-2.05---67# DOCA Rivermax89**Where to start:** This skill assumes DOCA is already installed,10**AND** that the NVIDIA Rivermax SDK is separately installed with a11valid Rivermax license present on the host. The agent's FIRST12action on any Rivermax question is to confirm both — without13Rivermax SDK + license, `doca-rmax` cannot function regardless14of how clean the DOCA-side code is, and this is the #1 first-app15confusion (DOCA does *not* bundle Rivermax; it wraps it). Open16[`TASKS.md`](TASKS.md) if the user wants to *do* something17(configure / build / modify / run / test / debug); open18[`CAPABILITIES.md`](CAPABILITIES.md) when the question is *what19can DOCA Rivermax express* on this version + this Rivermax20install. If the user has not installed DOCA yet, route to21[`doca-setup`](../../doca-setup/SKILL.md) first; for the Rivermax22SDK + license install itself, route to the public DOCA Rivermax23guide (slug `DOCA-Rivermax`) via24[`doca-public-knowledge-map`](../../doca-public-knowledge-map/SKILL.md).25If the user is asking *"how do I get packets to land on my26Rivermax input stream at all"*, the answer is layered:27`doca-rmax` is the *Rivermax integration* surface,28[`doca-eth`](../doca-eth/SKILL.md) is the *queue* surface that29carries the packets, and [`doca-flow`](../doca-flow/SKILL.md)30is the *steering* surface that directs them.3132## Example questions this skill answers well3334The CLASSES of DOCA Rivermax questions this skill is built to35answer, each with one worked example. The agent should treat the36*class* as the load-bearing piece — the worked example is a37single instance.3839- **"Can I even use `doca-rmax` on this host?"** — worked40 example: *"the public docs mention `doca-rmax`; is it41 usable without doing anything else?"*. Answered by the42 Rivermax-SDK + license precondition rule in43 [`CAPABILITIES.md ## Safety policy`](CAPABILITIES.md#safety-policy)44 + the env-prep checklist in45 [`TASKS.md ## configure`](TASKS.md#configure) step 1, which46 routes the install-side question to the public Rivermax guide47 via48 [`doca-public-knowledge-map`](../../doca-public-knowledge-map/SKILL.md)49 and refuses to recommend a fallback to `doca-eth` alone that50 would silently lose the timing properties.51- **"How do I set up a SMPTE ST 2110 video receive stream?"** —52 worked example: *"line up an inbound Rivermax stream on a53 representor of a BlueField port for first-run testing"*.54 Answered by the per-stream object lifecycle in55 [`TASKS.md ## configure`](TASKS.md#configure) +56 [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)57 input stream capability table.58- **"Which `doca_rmax_get_*_supported` query do I have to call before59 picking a PTP clock or hardware packet-placement order?"** — worked60 example: *"can this device + this Rivermax install use61 ST 2110-20 sequence-number placement?"*. Answered by the62 capability-query rule in63 [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)64 + step 3 in65 [`TASKS.md ## configure`](TASKS.md#configure).66- **"My stream is set up but no packets arrive — why?"** —67 worked example: *"the Rivermax stream object started cleanly68 but no recv events fire"*. Answered by the precondition matrix69 in70 [`CAPABILITIES.md ## Safety policy`](CAPABILITIES.md#safety-policy)71 + the env-prep checklist in72 [`TASKS.md ## configure`](TASKS.md#configure) step 1, which73 routes the steering side to74 [`doca-flow`](../doca-flow/SKILL.md), the queue side to75 [`doca-eth`](../doca-eth/SKILL.md), and the license side76 back to the Rivermax-side precondition.77- **"Is this Rivermax integration capability available on my78 device + my installed DOCA + my Rivermax SDK?"** — worked79 example: *"does this device + this Rivermax version advertise80 the PTP clock or placement mode I need"*. Answered by the capability-query81 rule in82 [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)83 + the version-and-device overlay in84 [`CAPABILITIES.md ## Version compatibility`](CAPABILITIES.md#version-compatibility),85 which adds the *Rivermax-side version is a second axis* rule86 on top of the canonical DOCA version-handling chain in87 [`doca-version`](../../doca-version/SKILL.md).88- **"What does this `DOCA_ERROR_*` from a Rivermax call mean and89 which layer caused it?"** — worked example:90 *"`DOCA_ERROR_NOT_SUPPORTED` from `doca_rmax_init()`"*.91 Answered by the Rivermax overlay on the92 cross-library taxonomy in93 [`CAPABILITIES.md ## Error taxonomy`](CAPABILITIES.md#error-taxonomy)94 + the layered ladder in95 [`TASKS.md ## debug`](TASKS.md#debug) that escalates to96 [`doca-debug`](../../doca-debug/SKILL.md), and which preserves97 the installed header's call-specific mapping: init-time98 `_NOT_SUPPORTED` routes first to Rivermax SDK / license checks;99 later errors are interpreted from the exact failing call rather100 than generalized into a license diagnosis.101102## Audience103104This skill serves **external developers building applications105that consume the DOCA Rivermax integration** — i.e., users whose106code calls `doca_rmax_*` (directly in C/C++, or through107FFI/bindings from another language) to drive timing-precise108media-over-IP streams (SMPTE ST 2110, real-time market data,109high-throughput scientific instrument streams) on top of a110separately-installed NVIDIA Rivermax SDK on a BlueField or111ConnectX host. It is *not* for NVIDIA developers contributing to112DOCA Rivermax itself, and it is *not* for users who want113best-effort packet I/O — for that, route to114[`doca-eth`](../doca-eth/SKILL.md) directly.115116**Language scope.** DOCA Rivermax ships as a C library with117`pkg-config` module name `doca-rmax`. The shipped samples118live under `/opt/mellanox/doca/samples/doca_rmax/` and are119written in C. C and C++ consumers are the canonical case; the120worked examples in `TASKS.md` assume that path. Other-language121consumers (Rust, Go, Python, …) consume the same `*.so` through122FFI or language-specific bindings; the skill's contribution in123that case is to keep the precondition rule (Rivermax SDK +124license), per-stream lifecycle, capability-discovery,125permission, scheduling-discipline, and error-taxonomy guidance126language-neutral, and to route the agent to the public C ABI as127the authoritative surface that any wrapper will eventually call.128129## When to load this skill130131Load this skill when the user is doing hands-on DOCA Rivermax132work, in any language. Concretely:133134- Confirming that the NVIDIA Rivermax SDK and a valid Rivermax135 license are present on the host before any DOCA-side code is136 written — this is a **mandatory precondition**, not an error137 path; without it the answer is *"`doca-rmax` cannot be138 used; pick a different library"*, not *"let's try and see what139 fails"*.140- Initializing the global DOCA Rivermax engine with141 `doca_rmax_init()` / `doca_rmax_release()`, then creating a142 `doca_rmax_in_stream` (receive) context with143 `doca_rmax_in_stream_create()` on a `doca_dev` opened against a144 physical port, a representor, or an SF, and converting it via145 `doca_rmax_in_stream_as_ctx()` before `doca_ctx_start()`. The146 public DOCA Rivermax API is receive-only — there is no147 transmit/output stream object.148- Configuring a Rivermax input stream for SMPTE ST 2110 audio +149 video over IP, real-time market data feeds, or a scientific150 instrument stream — with awareness that PTP-clock and151 hardware packet-placement-order support are device- and152 Rivermax-conditional and gate on the matching153 `doca_rmax_get_*_supported` query.154- Reading or setting Rivermax stream properties via155 `doca_rmax_in_stream_set_*` and querying device + Rivermax156 capability via the `doca_rmax_get_*_supported` family before157 assuming PTP-clock or hardware packet-placement-order support.158- Pairing the Rivermax stream with the `doca-eth` queue surface159 that carries the packets and the `doca-flow` rules that160 steer them — Rivermax does not program steering itself.161- Designing the real-time scheduling discipline for the162 streaming threads (the underlying Rivermax stack expects163 real-time priority for sub-microsecond jitter; the canonical164 scheduling guidance lives in the Rivermax SDK docs reachable165 through166 [`doca-public-knowledge-map`](../../doca-public-knowledge-map/SKILL.md)).167- Debugging a `DOCA_ERROR_*` returned from a Rivermax call168 (lifecycle vs. license / permission vs. capability vs.169 driver-below) where the cause may live in the DOCA-side170 wrapper, the underlying Rivermax stack, or the licensing171 layer.172- Designing or extending non-C bindings (Rust, Go, Python, …)173 that wrap the DOCA Rivermax C ABI — for the precondition174 rule, lifecycle, capability, permission, and scheduling rules175 the wrapper must honor.176177Do **not** load this skill for general DOCA orientation, for178installing DOCA itself, for installing the Rivermax SDK or179managing the Rivermax license file (those live in the public180Rivermax SDK guide reachable through181[`doca-public-knowledge-map`](../../doca-public-knowledge-map/SKILL.md)),182for best-effort packet I/O without timing requirements (use183[`doca-eth`](../doca-eth/SKILL.md) directly), or for pure184host-side data processing without networking. For DOCA185documentation orientation, use186[`doca-public-knowledge-map`](../../doca-public-knowledge-map/SKILL.md).187188## What this skill provides189190This is a **thin loader**. The body keeps only the orientation191needed to pick the right next file. The substantive192Rivermax-specific material lives in two companion files:193194- `CAPABILITIES.md` — what DOCA Rivermax can express on this195 version *and this Rivermax SDK install*: the196 Rivermax-as-hard-dependency rule, the receive-only197 `doca_rmax_in_stream` object model, the capability-query surface198 (`doca_rmax_get_*_supported`), the Rivermax error taxonomy199 (mapped onto the cross-library `DOCA_ERROR_*` set, with200 Rivermax-specific causes called out per row), the201 observability surface (per-stream progress engine events,202 capability snapshots, Rivermax-side license + driver state),203 and the safety policy that gates the Rivermax-SDK-present /204 license-present / device-access / scheduling-discipline205 preconditions.206- `TASKS.md` — step-by-step workflows for the six in-scope207 Rivermax verbs: `configure`, `build`, `modify`, `run`,208 `test`, `debug`. Plus a `Deferred task verbs` block that209 points out-of-scope questions (installing Rivermax,210 managing the license, programming steering, programming the211 underlying queue) at the right next skill, and a `Command212 appendix` of the recurring commands the agent reaches for.213214The skill assumes a host or BlueField where DOCA is already215installed at the standard location, the NVIDIA Rivermax SDK is216installed at its expected location with a valid license, and217the user has the privileges their public install profile218expects (typically sudo or `mlnx`-group membership to open a219`doca_dev` against a port). It does not cover installing DOCA —220that path goes through221[`doca-setup`](../../doca-setup/SKILL.md). It does not cover222installing Rivermax or its license — that path goes through the223public Rivermax SDK guide reachable through224[`doca-public-knowledge-map`](../../doca-public-knowledge-map/SKILL.md).225226## What this skill deliberately does not ship227228This skill is **agent guidance**, not a samples or templates229bundle. To keep the boundary clean, it deliberately does not230contain — and pull requests should not add:231232- **Pre-written DOCA Rivermax application source code, in any233 language.** The verified Rivermax source code is the shipped234 C samples at `/opt/mellanox/doca/samples/doca_rmax/<name>/`.235 The agent's job is to route the user to those files and236 prescribe a minimum-diff modification on them via the237 universal modify-a-sample workflow in238 [`doca-programming-guide`](../../doca-programming-guide/SKILL.md),239 layered with the Rivermax-specific overrides in240 [`TASKS.md ## modify`](TASKS.md#modify).241- **Standalone build manifests** (`meson.build`,242 `CMakeLists.txt`, `Cargo.toml`, …) parked inside the skill.243 The agent constructs the build manifest *in the user's244 project directory* against the user's installed DOCA, where245 `pkg-config --modversion doca-rmax` is the source of246 truth.247- **Quoted Rivermax symbol names beyond the public family248 pattern (`doca_rmax_get_*_supported` for capability discovery,249 receive-only `doca_rmax_in_stream` session objects driven by the250 standard DOCA Core context lifecycle).** Exact Rivermax251 symbol names are install-bound and Rivermax-SDK-version-252 bound; the agent should read them from the installed headers253 at $(pkg-config --variable=includedir doca-common) and from the254 public DOCA Rivermax guide rather than rely on agent memory.255- **A `samples/`, `bindings/`, or `reference/` subtree** of256 any kind. A mock or incomplete artifact in this skill's257 tree, even one labeled *"reference"*, is misleading: users258 will read it as buildable.259260## Loading order2612621. Read this `SKILL.md` first to confirm the user's question is263 in scope **and** that the Rivermax-SDK + license264 precondition has been considered.2652. **For the Rivermax capability matrix, the receive-only input266 stream model, capability-query rules, error taxonomy,267 observability, and safety policy, see268 [CAPABILITIES.md](CAPABILITIES.md).**2693. **For step-by-step workflows — configure, build, modify,270 run, test, debug — see [TASKS.md](TASKS.md).**271272Both companion files cross-link to each other,273[`doca-version`](../../doca-version/SKILL.md) for the canonical274DOCA version-handling rules (Rivermax adds a Rivermax-SDK-275version axis on top),276[`doca-eth`](../doca-eth/SKILL.md) for the queue surface277that carries the packets,278[`doca-flow`](../doca-flow/SKILL.md) for the steering side279that decides which packets land on which queue, and280[`doca-public-knowledge-map`](../../doca-public-knowledge-map/SKILL.md)281whenever the right answer is *"look it up in the public282Rivermax SDK guide or the installed package layout"* rather283than *"Rivermax-integration-specific guidance"*.284285## Related skills286287- [`doca-public-knowledge-map`](../../doca-public-knowledge-map/SKILL.md) —288 the routing table for every public DOCA documentation source289 and the on-disk layout of an installed DOCA package. The290 Rivermax URL slug is `DOCA-Rivermax`; the public Rivermax SDK291 guide (separate product) is reachable from the same routing292 table when the user asks how to install Rivermax or its293 license.294- [`doca-setup`](../../doca-setup/SKILL.md) — env preparation,295 install verification, port-state checks (`devlink dev show`,296 `ip link`), permission and group-membership requirements for297 opening a `doca_dev`. This skill assumes its preconditions298 are satisfied; the Rivermax-SDK + license preconditions are299 layered on top.300- [`doca-version`](../../doca-version/SKILL.md) — canonical301 DOCA version-handling rules. This skill's `## Version302 compatibility` cross-links the four-way match rule and adds303 the Rivermax-specific overlay (Rivermax SDK version is a304 second axis; the capability set on this host is the305 intersection of DOCA-side cap-query results and306 Rivermax-side capabilities).307- [`doca-structured-tools-contract`](../../doca-structured-tools-contract/SKILL.md) —308 the bundle's structured-tools precedence rule (detect /309 prefer / fall back / report). The Command appendix in310 [TASKS.md](TASKS.md) honors this contract.311- [`doca-programming-guide`](../../doca-programming-guide/SKILL.md) —312 general DOCA programming patterns shared by every library:313 the canonical `pkg-config` + meson build pattern, the314 universal modify-a-shipped-sample first-app workflow, the315 universal lifecycle, the cross-library `DOCA_ERROR_*`316 taxonomy, and the program-side debug order. This skill317 layers Rivermax specifics on top.318- [`doca-eth`](../doca-eth/SKILL.md) — the queue surface319 that carries the packets a Rivermax stream produces or320 consumes. DOCA Rivermax does *not* program the underlying321 queue itself; it integrates with the queue programmed via322 `doca-eth`. The two skills' lifecycles are independent.323- [`doca-flow`](../doca-flow/SKILL.md) — the steering324 surface that decides which packets land on which queue.325 DOCA Rivermax does *not* program steering itself; an empty326 Rivermax input stream almost always means a missing or wrong327 Flow rule (or a missing Rivermax license), not a Rivermax328 bug.329- [`doca-debug`](../../doca-debug/SKILL.md) — the330 cross-cutting debug ladder (install / version / build /331 link / runtime / program / driver). Rivermax-specific debug332 (license precondition gaps, stream-type / packet-rate333 capability mismatches, scheduling-discipline jitter334 symptoms) overlays on top of that ladder.335