메뉴
Creates minimal, reproducible test cases from bug reports to confirm the defect before and after a fix. Use when a bug is reported without a failing test, when the user needs a regression test for a fix, or when the user asks to reproduce a bug as a test.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
Generates code comments that explain non-obvious intent, constraints, and tradeoffs — not what the code already says. Use when code is correct but opaque, when documenting for future maintainers, or when a function's why is harder to see than its what.
Identifies recurring structural patterns in a codebase — idioms, copy-paste clones, homegrown abstractions — and characterizes each as a reusable template. Use when learning a codebase's conventions, when hunting for copy-paste that should be a function, or when documenting how this team does things.
Finds code by meaning, structure, or text across large codebases — picks the right search strategy (grep, AST query, call graph walk, semantic search) for the question being asked. Use when the user asks where something is implemented, when navigating unfamiliar code, or when a simple grep isn't enough.
Produces natural-language summaries of what code does at the function, class, module, or subsystem level, with length and abstraction scaled to the scope. Explains purpose, side effects, and non-obvious behavior rather than restating syntax. Use when onboarding to unfamiliar code, when the user asks what something does, when generating docstrings or architecture notes, or when preparing a handoff document.
Translates a single function or small code unit between programming languages, mapping idioms and preserving observable behavior. Use when porting one function, when the user pastes code and asks for it in another language, or as the per-unit primitive for larger migrations.
Identifies natural component boundaries inside a monolith by clustering the dependency graph, finding the cuts with minimum coupling. Use when planning to modularize or extract microservices, when deciding what can be deployed independently, or when the user asks where the seams in this codebase are.
SOC 직업 분류 기준
| name | bug-reproduction-test-generator |
| description | Creates minimal, reproducible test cases from bug reports to confirm the defect before and after a fix. Use when a bug is reported without a failing test, when the user needs a regression test for a fix, or when the user asks to reproduce a bug as a test. |
| license | Apache-2.0 |
| metadata | {"category":"debugging","suite":"general-secure-coding-agent-skills","version":"0.3.0","related":"bug-localization, test-case-reducer, bug-to-patch-generator"} |
A bug you can't reliably reproduce is a bug you can't reliably fix. This skill turns prose ("it crashes when I upload a big file") into an executable test that is red on the buggy code and will be green once it's fixed.
| Input available | First move |
|---|---|
| Stack trace | Top project-code frame → target function; exception → assertion |
| "Steps to reproduce" prose | Translate each step to a setup line; last step → action |
| Log excerpt | Find the first anomalous line; grep for the producer |
| Failing production request (curl, HAR) | Replay against a test harness; wrap in an assertion |
| Screenshot / "it looks wrong" | Not machine-checkable — ask for the data, not the UI |
| "It's intermittent" | Do not write a test yet — first stabilize (Step 4) |
Every reproduction test is (setup, action, assertion). Extract each:
Write a test that reproduces. It will be too big. Minimize mechanically:
| Dimension | Minimization |
|---|---|
| Setup state | Delete one setup line → still red? Keep deleting. Add back when it goes green. |
| Input size | Binary-bisect the input (half the file, half the list) until it goes green |
| Dependencies | Inline mocks; if removing a mock turns it green, that mock was load-bearing |
The target is: a test so small that when it fails, the fault is obvious from reading the test alone.
If the bug is non-deterministic, the test can't simply assert correct behavior — it'll pass by luck half the time.
| Source of flakiness | Stabilization |
|---|---|
| Timing / sleep-based | Replace wall-clock with an injected clock you advance manually |
| Thread interleaving | Force the bad interleaving with a latch/barrier; assert on the state you now deterministically reach |
| Network | Mock the transport; inject the specific response/error that triggers the bug |
| Test order | Run the test in isolation; if it passes alone, the bug is test pollution, not product code |
| Randomness | Seed the RNG with the seed that reproduces (if logged); if not, loop 1000× and assert all pass |
If you cannot stabilize, do not ship a test that retries. A retry loop in a reproduction test is an admission you don't understand the bug.
Report: "Exporting a CSV with a quote in a column value produces a file Excel can't open."
Triple extraction:
export_csv(rows)" → "" and field is quoted)Minimal test:
def test_csv_export_escapes_quotes():
rows = [{"name": 'say "hi"'}]
out = export_csv(rows)
assert out == 'name\r\n"say ""hi"""\r\n'
Three lines of meaning. When this fails, the reader immediately sees: quotes aren't escaped.
pytest.fail() after the call, with a timeout decorator. The test passes if it reaches fail() within the timeout.close()/__exit__ was called on the resource mock.requirement-enhancer.Once red: → bug-localization (test is the signal) → bug-to-patch-generator (test is the oracle).