Build, refine, test, validate, or list CLI-Anything harnesses for GUI applications in Cursor. Prefer slash commands /cli-anything, /cli-anything-refine, /cli-anything-test, /cli-anything-validate, and /cli-anything-list when available. This skill supports natural-language discovery of the generator; it does not replace those commands. Discovering already-published CLIs from CLI-Hub is a separate consumer skill (cli-hub-meta-skill).
57
66%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Passed
No findings from the security scan
Fix and improve this skill with Tessl
tessl review fix ./cursor-plugin/skills/cli-anything-generator/SKILL.mdThis skill is packaged inside the Cursor plugin as a supporting entry point. The primary generator UX is the plugin slash commands.
| Goal | Command |
|---|---|
| Full 7-phase build | /cli-anything <path-or-repo> |
| Expand coverage | /cli-anything-refine <path> [focus] |
| Run tests | /cli-anything-test <path> |
| Validate checklist | /cli-anything-validate <path> |
| List tools | /cli-anything-list |
If the user speaks naturally ("build a CLI for ./gimp"), run the same workflow as the matching command.
This skill is not the Hub consumer. For finding/installing published CLIs:
npx skills add HKUDS/CLI-Anything --skill cli-hub-meta-skill -g -yPer-app skills live under the repo skills/ tree and install via npx skills add.
PLUGIN_ROOT via discovery pointer then stamp:
~/.cursor/cli-anything-generator.root / %USERPROFILE%\.cursor\cli-anything-generator.root{install}/PLUGIN_ROOT.txt~/.cursor/plugins/local/cli-anything if references/HARNESS.md exists there{PLUGIN_ROOT}/references/HARNESS.md with an absolute path.{PLUGIN_ROOT}/references/commands/.{PLUGIN_ROOT}/references/guides/ on demand.{PLUGIN_ROOT}/scripts/repl_skin.py, preview_bundle.py,
skill_generator.py, templates/.| Document reference | Installed plugin path |
|---|---|
guides/... | references/guides/... |
cli-anything-plugin/repl_skin.py | scripts/repl_skin.py |
cli-anything-plugin/preview_bundle.py | scripts/preview_bundle.py |
cli-anything-plugin/skill_generator.py | scripts/skill_generator.py |
templates/SKILL.md.template | scripts/templates/SKILL.md.template |
docs/PREVIEW_PROTOCOL.md | references/docs/PREVIEW_PROTOCOL.md |
| Cursor tool | Role |
|---|---|
| Read | Methodology and source analysis |
| Grep / Glob | API and file discovery |
| Shell | Clone, pytest, pip, CLI invocation |
| Write / StrReplace | Harness implementation |
| Task | Optional parallel subtasks |
For build/refine/test/validate, maintain
{software-root}/agent-harness/.cli-anything-progress.json.
Create agent-harness/ when tracking starts. Do not advance without completing
the previous phase. Resume non-terminal runs.
Match Claude Code plugin / Codex skill output: same harness layout, namespace
packaging, --json, tests, and skill generation. Do not invent a different
harness format.
34f5195
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.