Menu
Run TinyTest tests and report results. Accepts a file, directory, or test name.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Deep-analyze a single PHP function — trace all code paths, find callers, classify bugs and smells, and generate exhaustive TinyTest unit tests. Useful before /test-cover-file when a function is complex.
Project-wide coverage audit — runs phpdbg across all tests, finds every uncovered function, analyzes each one in context with caller data, generates comprehensive tests for all code paths, and produces a structured bug/smell report with operator decisions. Requires phpdbg.
Project-wide bootstrap — scans all PHP source files, registers any without a test file in test_sources.md, and spawns sub-agents to generate initial test files for each. Run this first on a new project before using test-audit.
Achieve full coverage for a single PHP source file. Phase 1 iterates until every function is called. Phase 2 iterates until every reachable statement and branch is exercised. Requires phpdbg.
Diagnose and fix a failing TinyTest test. Reads the error output, compares test vs source, and repairs the test or explains the source bug.
Generate an initial TinyTest test file for a PHP source file. Creates the test file, runs it, fixes failures, and registers it in test_sources.md. Contains the full TinyTest assertion and annotation reference.
| name | test-run |
| description | Run TinyTest tests and report results. Accepts a file, directory, or test name. |
| argument-hint | [file|dir] [-t test_name] |
-f <file>-d <dir>-t <test_name> alongside -ftests/ directory and use -d tests/Find the tinytest path from AGENTS.md or CLAUDE.md in the project root.
php <tinytest_path>/tinytest.php -j -f <file>
If -j is not available (older tinytest), fall back to verbose mode:
php <tinytest_path>/tinytest.php -v -f <file>
From JSON output, report:
From verbose text output, look for:
err — these are failuresexpected [X] got [Y] — the assertion that failedFor each failure:
Report concisely: "X/Y tests passed. Z failures:" followed by a brief summary of each failure.