OpenSkillIndex ← back to search

My_harness validate-and-test

install

source · Clone the upstream repo

git clone https://github.com/WeiJun0507/my_harness

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/WeiJun0507/my_harness "$T" && mkdir -p ~/.claude/skills && cp -r "$T/.cursor/skills/validate-and-test" ~/.claude/skills/weijun0507-my-harness-validate-and-test && rm -rf "$T"

manifest: .cursor/skills/validate-and-test/SKILL.md

source content

Validate and Test

Choose the lightest validation that safely fits the task. Do not run heavyweight checks by default on a trivial change.

Default Order

Format changed files.
Run relevant lint or static analysis.
Write unit tests for new or changed behavior (
```
M
```
and above).
Run the tests written plus any relevant existing tests.
Do manual verification when automated tests are low value.
Escalate to broader regression checks for
```
L
```
and
```
XL
```
.

By Task Size

Format changed files if needed.
Run the smallest relevant static check.
Run tests only when the change is not purely mechanical.

Format changed files.
Run static analysis on the affected area.
Write unit tests for any new or changed behavior.
Run the new tests plus targeted existing tests.

Format changed files.
Run broader static analysis for the affected module or package.
Write unit tests for new and changed behavior.
Run those tests plus nearby regression checks.
Verify shared callers, contracts, or user flows touched by the change.

XL

Validate by phase.
Write tests for each phase before moving to the next.
Use broader regression coverage.
Re-check assumptions after each phase.
Treat unresolved risk as a blocker, not a footnote.

Output Format

Format: PASS / FAIL
Static checks: PASS / FAIL
Tests written: YES / NO / NOT APPLICABLE
Tests result: PASS / FAIL / NOT RUN
Manual verification: PASS / FAIL / NOT RUN
Residual risk: none / [details]
Result: PASS / BLOCKED