My little collection of plugins: opencode-agent-memory, opencode-agent-skills, opencode-beads, and opencode-handoff #61
joshuadavidthomas
started this conversation in
Plugin
Replies: 1 comment
|
Good job! We built something similar internally, using very similar patterns, we have few scattered plugins. I'm interested in working to build a more complete community driven solution |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
I've been building a few plugins for OpenCode and wanted to share them in case they're useful to anyone else.
opencode-agent-memory - Gives the agent persistent, self-editable memory blocks inspired by Letta agents. Basically
AGENTS.mdwith a harness: structured blocks with metadata and prompts injected early into the context that encourages the agent to actively maintain its own memory, with a lot of eye-roll enducing philosophical framing in the prompts borrowed from Letta (and yet somehow it does seem to work!).opencode-agent-skills - Dynamic skills loader that discovers skills from project, user, and plugin directories. Compatible with existing Claude Code skills and optionally integrates with Superpowers (https://github.com/obra/superpowers). Different from opencode-skills as it only has 4 tools for reading skills and executing skill scripts, whereas that plugin creates a tool per skill.
opencode-beads - Integration for Steve Yegge's beads issue tracker. Injects beads context automatically on session start and after compaction, exposes
/bd-*commands, and includes a task agent for autonomous issue completion.opencode-handoff - Creates focused handoff prompts for continuing work in a new session. Inspired by Amp's
/handoffcommand. Generates a continuation prompt with relevant file references and opens a new session with it as an editable draft.All reactions