workflow

Original：🇨🇳 Chinese

Translated

3 scripts

A thin orchestration layer for embeddedskills, used to discover projects in the current workspace, select build/flash/debug/observe backends, connect .embeddedskills/state.json, and aggregate results from underlying skills. Triggered when the user explicitly enters one of the following commands: "One-click Build & Flash", "Auto Diagnose", "Chain build -> flash -> debug -> observe", or explicitly calls /workflow.

2installs

Sourcezhinkgit/embeddedskills

Added on2026-05-19

NPX Install

npx skill4agent add zhinkgit/embeddedskills workflow

SKILL.md Content (Chinese)

View Translation Comparison →

Workflow Orchestration Layer

This skill does not reimplement underlying logic, only handles discovery, selection, chaining, and aggregation.

The

observe

phase currently provides candidate backends including

jlink:rtt

,

jlink:swo

,

openocd:semihosting

,

openocd:itm

,

probe-rs:rtt

.

Commands

bash

python <skill-dir>/scripts/workflow_plan.py --json
python <skill-dir>/scripts/workflow_run.py plan --json
python <skill-dir>/scripts/workflow_run.py build --json
python <skill-dir>/scripts/workflow_run.py build-flash --json
python <skill-dir>/scripts/workflow_run.py build-debug --json
python <skill-dir>/scripts/workflow_run.py observe --json
python <skill-dir>/scripts/workflow_run.py diagnose --json

Configuration Instructions

Workflow no longer maintains an independent project configuration structure; all project parameters are read uniformly from

.embeddedskills/config.json

.

Configuration Structure

The

workflow

section in

.embeddedskills/config.json

only contains preferred backend configurations:

json

{
  "workflow": {
    "preferred_build": "auto",
    "preferred_flash": "auto",
    "preferred_debug": "auto",
    "preferred_observe": "auto"
  }
}

Workflow obtains project parameters (such as

keil.project

,

jlink.device

,

probe-rs.chip

, etc.) by reading configuration sections of other skills in

.embeddedskills/config.json

.

Parameter Parsing Order

Judge in sequence according to the following decision tree, stop once a match is found:

CLI Parameters (highest priority)
- Condition: The user passes parameters like
```
--build-backend
```
  ,
```
--flash-backend
```
  via command line
- Example:
```
workflow_run.py build-flash --build-backend=keil --flash-backend=jlink
```
- Behavior: Directly use the backend specified by the parameter, skip subsequent steps
Configuration File (second priority)
- Condition: Not specified via CLI, and the corresponding
```
preferred_*
```
  field in the
```
workflow
```
  section of
```
.embeddedskills/config.json
```
  is not
```
"auto"
```
- Example:
```
"preferred_build": "keil"
```
  → Use keil as the build backend
- Behavior: Read the configuration value and use it, skip automatic discovery
Automatic Discovery (fallback)
- Condition: Not specified via CLI, and
```
preferred_*
```
  in the configuration is
```
"auto"
```
  or the field is missing
- Example:
```
"preferred_flash": "auto"
```
  → Scan the workspace to automatically infer available flash backends
- Behavior: Enumerate the list of candidate backends; if there is only one, use it directly; if there are multiple, return the list for user confirmation

After successful execution, the actually used backend will be automatically written back to the

workflow

section of

.embeddedskills/config.json

.

Rules

When multiple projects or multiple candidate backends are discovered, only return the candidate list without automatic guessing
Prioritize connecting build, flash, debug, and observe via
```
.embeddedskills/state.json
```
```
observe
```
only generates recommended commands, and does not directly occupy the observation channel for a long time within the workflow
When a failure occurs, prioritize returning which phase failed and the structured error of the underlying script
Collaboration between workflow and other Skills only happens via
```
.embeddedskills/config.json
```
,
```
.embeddedskills/state.json
```
, and subprocess calls to underlying Skills