| name | chip-tool-testing |
| description | Guidelines for building Matter examples and chip-tool, and testing examples using chip-tool. |
Testing Matter Examples with Chip-Tool
This skill provides guidelines for building Matter example applications and
chip-tool, and using chip-tool to commission and interact with the examples
in a simulated environment (like Linux POSIX).
[!NOTE] While chip-tool and examples support multiple platforms, these
guidelines focus on the Linux/POSIX environment, which is the standard
environment for agent execution.
Building from Source
Always build the chip-tool and the Matter example application from the same
revision of the repository to ensure compatibility.
Building chip-tool
Use the build script to build chip-tool for Linux:
source scripts/activate.sh
./scripts/build/build_examples.py --target linux-x64-chip-tool-clang build
The binary will be located in out/linux-x64-chip-tool-clang/chip-tool.
Building Examples
Use the build script to build examples (e.g., all-devices-app):
source scripts/activate.sh
./scripts/build/build_examples.py --target linux-x64-all-devices-boringssl build
The binary will be located in
out/linux-x64-all-devices-boringssl/all-devices-app.
Testing Guidelines
Running the Example
Run the example application (usually in a separate terminal or the background):
./out/linux-x64-all-devices-boringssl/all-devices-app
Note: Use a clean KVS file (by specifying a new, non-existent file path with
--KVS) or clear the existing one (by deleting the file) to ensure the
application enters commissioning mode on startup.
1. Determine Pairing Credentials
Check the application logs on startup to find the setup PIN code and
discriminator. Example log:
[DL] Setup Pin Code: 20202021
[DL] Setup Discriminator: 3840 (0xF00)
[!NOTE] The discriminator value 3840 (0xF00) is the standard default for
many Matter examples when not explicitly configured.
2. Commissioning (Pairing)
Make sure the device is in commissioning mode. If using a persistent storage
(KVS), you may need to use a clean KVS file or clear it to force commissioning
mode on startup. On Linux, the default KVS file is /tmp/chip_kvs:
rm /tmp/chip_kvs
[!NOTE] > Persistence Testing: If you are performing a persistence test
(verifying state across restarts), do NOT delete the KVS file. Reuse the same
KVS file to preserve stored attributes.
Over IP (On-Network)
If the device is already on the same IP network (e.g., running locally on the
same host):
-
With Long Discriminator:
chip-tool pairing onnetwork-long <node_id> <setup_pin> <discriminator>
-
Without Discriminator (looks for any commissionable device):
chip-tool pairing onnetwork <node_id> <setup_pin>
-
With Manual Pairing Code or QR Code:
chip-tool pairing code <node_id> <pairing_code_or_qrcode>
-
Direct IP and Port (useful if mDNS resolution fails):
chip-tool pairing already-discovered <node_id> <setup_pin> <device_ip> <device_port>
[!CAUTION] If you encounter attestation failures in simulated
environments, you may need to add --bypass-attestation-verifier true to
the command. Use this flag ONLY for local simulated testing when you
explicitly accept the security tradeoff of bypassing device attestation.
3. Interacting with Clusters
After successful commissioning, you can use chip-tool to interact with data
model clusters.
-
Read Attribute:
chip-tool <cluster_name> read <attribute_name> <node_id> <endpoint_id>
Example: chip-tool basicinformation read vendor-id 0x8016 0
-
Invoke Command:
chip-tool <cluster_name> <command_name> [arguments] <node_id> <endpoint_id>
Example: chip-tool onoff toggle 0x8016 1
Automated YAML Testing with run_test_suite.py
Instead of starting the application and running chip-tool commands manually,
you can run automated YAML tests using the run_test_suite.py script. The
script manages launching the example application, commissioning, running the
test steps, and cleanup.
Environment Activation
Make sure you are in the python virtual environment:
source out/venv/bin/activate
export PW_ACTIVATE_SKIP_CHECKS=1
Running a Test
Run a YAML test by specifying the target test name (without the .yaml
extension), the chip-tool python wrapper, and the target application path:
./scripts/tests/run_test_suite.py --runner chip_tool_python --target <test_name> run --tool-path chip-tool-with-python:scripts/tests/chipyaml/chiptool.py --app-path <app_key>:<binary_path>
Example: Testing Humidity Sensor in all-devices-app
If the YAML test (e.g. Test_TC_RH_2_1.yaml) specifies a dynamic CI app
target (such as all-devices with custom arguments), you must map that target
key to the compiled binary path:
./scripts/tests/run_test_suite.py --runner chip_tool_python --target Test_TC_RH_2_1 run --tool-path chip-tool-with-python:scripts/tests/chipyaml/chiptool.py --app-path all-devices:out/linux-x64-all-devices-clang/all-devices-app
Troubleshooting Automated Runs
- Missing tool key 'chip-tool-with-python': You must explicitly pass
--tool-path chip-tool-with-python:scripts/tests/chipyaml/chiptool.py to
the run subcommand.
- Missing application path: Ensure that any application specified in the
test's
CI block (or all-clusters if no CI block is present) is
explicitly mapped with --app-path.
Using Chip-Tool Help
chip-tool has a built-in help system that allows you to discover available
clusters and commands. Running any command without all required parameters will
show a help screen.
This interactive help is useful for exploring the data model and finding the
exact syntax needed.
Troubleshooting
- Timeout: If a command times out, you can increase it with
--timeout <seconds>.
- Configuration Cache:
chip-tool caches state in chip_tool_config.ini
(often in the current directory or /tmp). Deleting this file can resolve
stale configuration issues. You can also specify a custom directory with
--storage-directory <path>.
Reporting Results
When completing a test task using this skill, you should output a structured
report to the user. Since every test may verify different features, do not force
a fixed sequence of steps. Instead, follow these general guidelines for
reporting each item you verified:
For each verified feature, command, or configuration item, provide:
- What was verified: A clear description of the intent (e.g., "Verified
that Endpoint 2 is a child of Endpoint 1").
- The command ran: The exact
chip-tool or shell command you used.
- The proof (logs): A snippet of the output or logs that proves the
verification was successful (e.g., the attribute value read or the success
status).
General Guidelines for Content:
- Build Process: Mention which targets were built and if any custom flags
were used.
- Startup: Include the command used to run the application, and the
pairing credentials (Setup PIN and Discriminator) found in the logs.
- Commissioning: Show the command used and the log snippet confirming
success.
- Data Model: If hierarchy or device composition was verified, show the
relevant
Descriptor reads.
- Cluster Interaction: For any attribute read/write or command invocation,
show the command and the response status.