skills: add Zephyr agent skills#1
Conversation
2706fcb to
4e30f74
Compare
|
@codex review this PR thoroughy |
|
Codex Review: Didn't find any major issues. Keep it up! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
4e30f74 to
1793946
Compare
|
@codex review again |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 17939468e2
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
1793946 to
96f54c1
Compare
|
@codex review again |
Helps agents reproduce and isolate Zephyr build, boot, board, and runtime failures. Signed-off-by: Chao Liu <[email protected]>
Helps agents investigate Zephyr source, history, and constraints before making changes. Signed-off-by: Chao Liu <[email protected]>
Guides agents in reviewing Zephyr patches for correctness, coverage, ownership, and merge readiness. Signed-off-by: Chao Liu <[email protected]>
Guides agents in planning and reviewing Zephyr safety evidence, traceability, and certification gaps. Signed-off-by: Chao Liu <[email protected]>
96f54c1 to
2f710ab
Compare
|
Is it ready now? |
Yes. |
2 participants
Summary
review, and safety evidence work.
.zephyr-skills/<agent-task>/, separate from Zephyr build outputs.zephyr-debugwith Zephyr logging, QEMU GDB stubs, OpenOCD, J-Link,scripted GDB sessions, and optional Tmux or Zellij orchestration.
agent client.
Why
The complete skill set has been tested with GPT-5.5 and GPT-5.6. Based on the
performance of these newer models, common debugging and build/test workflows
are already internalized and perform well without dedicated skill
instructions. We therefore do not duplicate those generic capabilities as
standalone skills, because overly prescriptive instructions can unnecessarily
constrain model behavior.
The remaining skills focus on Zephyr-specific context, decision boundaries,
auditability, and specialized workflows that still benefit from explicit
guidance.
Validation
quick_validate.py.npx skillswithout specifying anagent.
bundled resources.
zephyr-build-testskill is no longer discoverable.git diff --check origin/main..HEAD.