Use when adding or updating Go CLI E2E coverage for one `tests/cli_e2e/{domain}` domain of the compiled `lark-cli`, especially when the work requires live `--help` or `schema` exploration, scenario-based `clie2e.RunCmd` workflows, and per-domain `coverage.md` maintenance.
SkillScore68Maintained
Automatic
◆QA & Test Engineering pack
Last verified 8/7/2026
What it changes
Passed 13 of 19 test cases with the skill, and 8 of 19 without it.
Without skill40%
With skill68%
Checks it passed that the model alone missed
Places URL parameters in Params
Asserts code: 0 for API responses
Places request payload JSON in Data
and 12 more in the full report
11 fixed6 passed either way4 still failing4 broke
Pass-rate lift
+28 pp
Tokens
+11%
Agent turns
0%
Response time
-58%
Model gemini-3.6-flashMeasured 8/7/2026v117 of 25 cases passed with skill · 19 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 10 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 10
Prompt
Before updating assertions on a modified CLI command output, what step should be taken if field names or JSON shapes are uncertain?
What was checked, without and with the skill
Checked
Without skill
With skill
Inspects CLI schema or help output using lark-cli schema or --help
Missed
Passed
case-23 · gemini-3.6-flash · 8/7/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md114 linessingle file
CLI E2E Testcase Writer
Work on one domain per run. Produce exactly two artifacts for that domain:
workflow testcase files under tests/cli_e2e/{domain}/
tests/cli_e2e/{domain}/coverage.md
Focus on domain testcase files. Do not change shared E2E support code such as tests/cli_e2e/core.go unless the user explicitly asks. Treat tests/cli_e2e/demo/ as reference only.
Core standard
Make the testcase scenario-based and self-contained.
Prove one workflow end to end: create plus follow-up read, or mutate plus teardown.
Prefer one file per workflow or one closely related feature.
For mutable flows, prove persisted state with read-after-write assertions, not just exit code.
Leave prerequisite-heavy paths uncovered when they cannot be proven, and explain why in coverage.md.
A leaf command is one that executes an action — it has no further subcommands.
If lark-cli <domain> <group> --help lists no subcommands, <group> itself is the leaf.
Count task +create as one leaf and task tasks get as one leaf.
Do not count parameter combinations.
Reuse coverage already present under tests/cli_e2e/{domain}/. Do not count tests/cli_e2e/demo/.
3. Choose the proof surface before editing
Identify the provable risks for the touched workflow: invalid input, missing prerequisite, identity or permission, state transition, output shape, cleanup safety. If only the happy path is testable, document the blocked risk areas in coverage.md.