Use when the user wants Codex to build, refine, test, validate, or list CLI-Anything harnesses for GUI applications or source repositories. Adapts the full CLI-Anything methodology to Codex without changing the generated Python harness format.
77
—
Does it follow best practices?
Impact
95%
1.55xAverage score across 3 eval scenarios
Passed
No known issues
Harness directory structure and packaging
Top-level layout
50%
100%
SOFTWARE.md present
0%
100%
Namespace __init__.py absent
0%
100%
Package __init__.py present
0%
100%
find_namespace_packages
50%
100%
console_scripts entry point
0%
100%
Click subcommands
100%
100%
--json flag
100%
100%
REPL mode present
100%
100%
ReplSkin copied to utils
0%
100%
Required inner files
12%
100%
build_report target
100%
100%
build_report files list
100%
100%
build_report risks
100%
100%
Test suite conventions and structure
TEST.md exists
0%
100%
TEST.md before test code
0%
28%
test_core.py exists
0%
100%
test_full_e2e.py exists
0%
100%
test_core is unit-level
100%
100%
e2e uses real files
100%
100%
Programmatic output verification
100%
100%
_resolve_cli usage
0%
100%
CLI_ANYTHING_FORCE_INSTALLED
0%
100%
Installed command invoked
100%
100%
TEST.md content
0%
100%
test_notes output verification
100%
100%
Session state locking and dry-run
session.py exists
50%
50%
File locking used
100%
100%
Lock covers write
100%
100%
Undo/redo tracking
100%
100%
Auto-save on mutation
100%
100%
--dry-run flag present
100%
100%
Dry-run skips side effects
100%
30%
Locking mechanism named
100%
100%
Undo/redo explained
100%
100%
Dry-run explained
100%
100%
Limitations section
100%
100%
dc73924
Table of Contents
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.