Hi all, Following up on James's request for early versions, I'd be happy to put together the testing and database skills, building on the AGENTS.md work in https://github.com/apache/fineract/pull/6642. *Testing:* Choosing the right test suite, using Feign for new integration tests, handling business dates, keeping tests isolated, and verifying changes. *Database:* Understanding Liquibase parts and includes, PostgreSQL/MySQL contexts, DDL safety and backward-compatibility checks, and seed rows.
I'll keep both skills short and practical, with step-by-step guidance linked to our existing docs or the relevant CI checks. I'll start with testing and then move on to database, placing them under .agents/skills/ to follow Apache Fory's approach, unless anyone prefers skills/. Regards, Abhishek On Mon, 19 Oct 2026 at 01:31, Edward Kang <[email protected]> wrote: > +1 > > And thanks for the discussion, everyone. Really insightful points that has > shaped my mind on a lot of things. > > On Sun, Oct 18, 2026 at 12:54 PM VICTOR MANUEL ROMERO RODRIGUEZ < > [email protected]> wrote: > >> +1 >> >> El dom, 18 oct 2026 a las 13:06, Ádám Sághy (<[email protected]>) >> escribió: >> >>> Hi >>> >>> I think it is a great idea. >>> >>> Regards, >>> Adam S >>> >>> On Oct 18, 2026, at 7:31 PM, James Dailey <[email protected]> wrote: >>> >>> I'd like to come back to this and to recommend that we move forward with >>> Paul's (Tofunmi's) recommendations: >>> >>> 1. skills/code-review/SKILL.md >>> 2. skills/database/SKILL.md >>> 3. skills/cob-batch/SKILL.md >>> 4. skills/testing/SKILL.md (For example: New integration tests MUST >>> use the Feign-based Fineract client (fineract-client-feign), NOT >>> RestAssured) >>> 5. skills/security/SKILL.md >>> >>> The discussions I had at Community Over Code and the sessions I attended >>> strongly indicated that we need to start to take these steps. The content >>> of those skills needs to be reviewed by Committers and that's where some of >>> the value of the community lies. >>> >>> I think those are just project files that are the functional equivalent >>> of "coding guidelines". >>> >>> I would like to see an early version of these, minimally designed and >>> tightly related to our project norms as we know them today. Get those into >>> the review process. >>> >>> We're not saying - "everyone should use AI without control", we're >>> saying, if you are pointing your AI tooling at our project, it and you >>> should know about how we do things. >>> >>> Thanks >>> >>> On Fri, Sep 4, 2026 at 10:41 AM Adam Monsen <[email protected]> wrote: >>> >>>> Thanks for the links, Aman. Fascinating. I hadn't considered how >>>> requiring, e.g. "Assisted-By:" is basically free advertising for AI >>>> vendors. >>>> >>>> My knowledge & experience here is limited but I still don't see any >>>> blockers, so I moved my proposal >>>> <https://github.com/apache/fineract/pull/6366> back to "ready for >>>> review". Reviewers, please focus on the new AI Policy section in >>>> CONTRIBUTING.md >>>> <https://github.com/apache/fineract/pull/6366/changes#diff-eca12c0a30e25b4b46522ebf89465a03ba72a03f540796c979137931d8f92055R275>. >>>> That's the only relevant change. Comments/thumbs-ups/thumbs-downs welcome >>>> (add your 👍 or 👎 or comments to the PR >>>> <https://github.com/apache/fineract/pull/6366/>, or keep this dev list >>>> thread going). I'll check back in on this in 72 hrs. If the discussion is >>>> over and I get an approval I'll go ahead and merge. If there's a veto, I'll >>>> pause my work. >>>> >>> >>> > > -- > Edward E. Kang > [email protected] > 972-768-6940 <(972)%20768-6940> >
