AI Agent tooling update: new CLI features and more capable Hive skill

In my latest post I said I’d post more often, so here we go with a proper update :)

What landed this week in hive-tx-cli

This week we merged two PRs in hive-tx-cli (kudos to Hive account@jarvie and Hive account@hivetrending). Big step forward in terms of features and command line options.

The first PR (by Hive account@hivetrending) added support for publishing post with burn rewards and beneficiaries, check account RC and verify (aka wait) for a broadcasted transaction to be included in a block.

The second one (by Hive account@jarvie) landed a few days later and brought in a lot of the missing command coverage. Publishing flows are now more polished, social/community actions are easier to run, and account-side operations are better represented. On top of that, query capabilities expanded in practical ways, especially around balance/RC/feed/replies and URL-based lookups.

The repository README is also up to date with the latest changes.

Repository: https://github.com/asgarth/hive-tx-cli

A more capable hive skill

With all the changes above, the hive skill needed a refresh to properly support all the new capabilities.

The guidance is now clearer around posting vs active key usage, examples are more practical (including URL-based flows), and the publishing metadata section is more complete, especially for AI-assisted workflows. Troubleshooting notes were also tightened up so it behaves better in automation/script contexts.

Skill path: https://github.com/asgarth/skills/tree/main/skills/hive

New skill: venice-ai

We also introduced a brand new venice-ai skill, and this one opens up a lot for media-heavy agent workflows. If you're not familiar with it, Venice.ai is one of the providers we use on PeakD for AI image generation. Check it out if you haven't tried it yet, it's free: https://peakd.com/ai 😉

It includes ready-to-use scripts and references for image generation, image editing/inpainting, upscaling/enhancement, background removal, and video generation with a quote/confirmation flow before spending.

A follow-up update adjusted defaults so watermark-free output is the standard path in generation examples/scripts (with explicit opt-in behavior when needed), while keeping cost-effective model defaults.

Skill path: https://github.com/asgarth/skills/tree/main/skills/venice-ai

Quick examples

# CLI
hive publish --permlink my-post --title "My Post" --body-file ./post.md --metadata '{"app":"hive-tx-cli"}'
hive vote --url https://peakd.com/@author/permlink --weight 100 --wait
hive community subscribe hive-12345

# Skills
npx skills add https://github.com/asgarth/skills --skill hive
npx skills add https://github.com/asgarth/skills --skill venice-ai

Wrap-up

Overall: the CLI got a substantial expansion this week, the hive skill now reflects that reality much better, and venice-ai gives us a clean, reusable media pipeline skill.

If you’re building agent flows on Hive, this should remove a lot of repetitive setup and glue code.