Kianoosh Raika
@djKianoosh
- Virginia
- github.com/djKianoosh
- Joined October 2026
Commit activity
24 commits in the last year
longest streak 3 dayscurrent 1
665 / 1,000 XP335 XP to wheel
Yesterday
Tue, Nov 17
replying to @IsaacHatilima
I love skills and I think bespoke skills are super important and powerful.
now, it's definitely nice for newcomers or someone new to a team/project to have some existing skills, especially if the codebase owner wants to enforce or otherwise nudge others towards something…
replying to @phpete1
I used the open source version, self hosted everything. the runner runs every hour, and projects opt in by adding a gitlab topic to their repo and a renovate.json config. I created an AI skill for devs to help them onboard, with a bunch of presets for configs that are kind of…
replying to @phpete1
just turned it on for an org with over 5000 repos over many many teams. it's great. veeerrry flexible and everyone can configure it to line up with their own team cadence and capacity.
Tue, Nov 3
replying to @slogsdon
Here's one possibility with just 4:
- PM with tech background, business analyst, knows how to work stakeholders
- UI/UX/devops/automation with backend experience, vision, knows how to deliver
- 2 data/backend/automation devs, know how to scale/orchestrate data pipelines and get…
Wed, Oct 28
Thu, Oct 22
replying to @nightcoding
It's not even that beefy of a linux server. Good amount of memory and large disks but nothing crazy special. Antigravity and google take care of the rest. If I had a new mac mini I could take advantage of their local foundation model though. A remote accessible harness like agy…
what is a good amount of memory? 😄 I have a spare Lenovo m720q with 16gb of ram which just collects dust in the moment. Suitable for it?
Wed, Oct 21
replying to @nightcoding
antigravity with remote control is pretty awesome tbh. I run the agy cli on my linux server at home and can reach it from any browser.
replying to @jrecyclebin
you can also create functions (they call them modules). this one i use all the time to represent any json i come across as a schema:
https://gitlab.com/-/snippets/4884698

GitLabjq module : json schema inference ($4884698) · Snippets · GitLabGitLab.com
Sun, Oct 18
replying to @ricki-io
i am still a bottleneck even though I make AI use some skills to do the reviews and
glab cli or gh to submit the actual comment. I still "steer" it with my own commentary in the prompt.
It's a matter of trust and "taste" (yeah overloaded term these days) really.replying to @djKianoosh
I started using "phase" "modality" and "domain" as a way to classify. Phase as in, what phase of work or when you reach for it, modality is the how, and domain is the "where" you'd use it.…
GitHubskills-enzebot/skill-architect/references/taxonomy-guide.md at main · redsash/skills-enzebotAn AI Skills Forge. Contribute to redsash/skills-enzebot development by creating an account on GitHub.
That's @djKianoosh's first reply.