docs(amdsmi): add pressure-testing-skills skill and mechanizable skill checks#8913
Open
marifamd wants to merge 3 commits into
Open
docs(amdsmi): add pressure-testing-skills skill and mechanizable skill checks#8913marifamd wants to merge 3 commits into
marifamd wants to merge 3 commits into
Conversation
Add a skill for validating a SKILL.md, prompt, or rule by running a fresh subagent that follows the doc literally on a known-answer fixture, capturing friction, and iterating to determinism before minimizing. Includes lessons learned from pressure-testing other skills: give a command instead of a prose-only rule, keep shortcut greps in sync with what the doc reads, define a rule's trigger so it is not over-applied, and never grade against a truncated fixture. Cross-references the writing-skills skill. Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
- Define what triggers the cascade (new/changed function) and give a table for non-function edits (enum, struct, bug fix, comment) that do not all cascade the same way. - Note the Quick Check greps only the five name-bearing layers; call out generator.py (a parser, verify by regenerating) and docs/ (prose) as not covered, so a clean grep is not mistaken for full coverage. - Add rust-interface and goamdsmi_shim as cascade consumers and grep targets; both bind amdsmi_* directly and were previously omitted. Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
- Add a copy-paste title regex plus length test so the Conventional Commits
structure and 72-char cap are self-checkable, not just prose.
- Clarify the imperative-mood rule to reject noun-phrase subjects
("More cleanup", "status-string tests"), with a Common Mistakes row.
Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
✅ All Policy Checks Passed
📖 Need help? See the Policy FAQ for details on every check and how to fix failures. |
|
🚫 Please fix the failed policies before requesting reviews. The following policy checks failed:
The |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Motivation
Consolidates three agent-documentation improvements produced while pressure-testing amd-smi's skills into a single reviewable change. Each commit is one idea.
Technical Details
pressure-testing-skillsskill. Validates a SKILL.md, prompt, or rule by running a fresh subagent that follows the doc literally on a known-answer fixture, captures friction, and iterates to determinism before minimizing. Cross-referenceswriting-skills.project-layout.md. Defines what triggers the cascade (new or changed function), adds a table for non-function edits, notes that the Quick Check grep only covers the five name-bearing layers, and addsrust-interfaceandgoamdsmi_shimas cascade consumers.Docs-only: touches
.claude/skill and rule files, with no library, CLI, or test code changed.Issue Tracking
JIRA ID : AILITOOLS-278
Test Plan
Not applicable to runtime. The
pressure-testing-skillsskill itself is the validation method: each edited skill or rule was exercised by a fresh subagent on a fixture during development.Test Result
Skill and rule docs render correctly, and no code paths are affected.
Submission Checklist
.claude/agent files, with no code or ABI impact.