99 in small chunks that are easy for a human to digest.
1010
1111 No 'vibe' coded contributions where no human understands what was produced
12- or how it works will be accepted.
12+ or how it works will be accepted. Such things are fine for rapid
13+ experimentation, just not final product.
1314
14- Can developers use AI/LLM tools to contribute to Test2-Harness?
15+ These rules exist to serve one goal: the humans who maintain this code
16+ later must be able to understand it. Proven contributors have some leeway
17+ where that goal is still met.
1518
16- Developers who contribute to Test2-Harness are allowed to use whatever
17- tools they decide are necessary for their work. The project will not
18- micromanage a contributor's workflow. This includes AI/LLM tools.
19+ Who these rules are for, and how strict they are
20+
21+ The bar every contribution must clear is that the humans maintaining the
22+ code going forward can understand it. The rest of this policy is a guide
23+ for getting new contributors and new contributions over that bar. It is
24+ not an unbreakable law.
25+
26+ Active project members with a track record of quality work have more
27+ freedom. They may deviate from the specifics below — how work is chunked,
28+ whether corrections land as their own commit, how AI use is noted — as
29+ long as what they contribute stays high quality and understandable to the
30+ people who will maintain it.
31+
32+ How much flexibility applies depends on the contribution itself and on
33+ proven contributors being willing to do the review. A larger or looser
34+ contribution is acceptable when someone trusted has actually read it and
35+ stands behind it. When that is not the case, follow the rules as written.
36+
37+ Can developers use AI/LLM tools to contribute?
38+
39+ Developers who contribute are allowed to use whatever tools they decide are
40+ necessary for their work. The project will not micromanage a contributor's
41+ workflow. This includes AI/LLM tools.
1942
2043AI/LLM code MUST be human vetted
2144
22- Any code that goes into Test2::Harness should be vetted by a human.
23- Maintainers who merge pull requests MUST verify the accuracy and utility of
24- any PR.
45+ Any code that goes in should be vetted by a human. Maintainers who merge
46+ pull requests MUST verify the accuracy and utility of any PR.
2547
2648 If the AI/LLM generated code needs additional edits or corrections
2749 before/during merge, those MUST be a separate commit indicating what human
@@ -42,6 +64,13 @@ No 'vibe' coding
4264 assist, and be used as a tool. But it cannot do all the work, and it cannot
4365 replace your own understanding of the code you are contributing.
4466
67+ Documentation follows the project's documentation policy
68+
69+ Documentation may be AI written, and may even be identifiably so, but it
70+ must meet the same readability and content standards as human-written
71+ documentation. It must be concise, natural to read, useful to its audience,
72+ and free of needless repetition.
73+
4574Significant AI/LLM code MUST be noted as such.
4675
4776 Code from an LLM/AI should be noted as such, either in the commit, the
@@ -66,10 +95,10 @@ Significant AI/LLM code MUST be noted as such.
6695
6796 The reason for noting it is so that we know if there is a human who
6897 understands it that can be asked questions about the code later. Such a
69- human does exist when AI is an assistant, it does not exit when AI is the
98+ human does exist when AI is an assistant, it does not exist when AI is the
7099 primary author.
71100
72- This is entirly due to maintaner preference, it is not intended for an
101+ This is entirely due to maintainer preference, it is not intended for an
73102 exhaustive audit or to single-out contributors who use or reject AI
74103 assistance. Using this rule as an excuse to troll or attack a contributor
75104 will not be tolerated. If you think someone missed the mark on noting AI
0 commit comments