proven pulsar
Hey guys 👋
I’ve been building Forge, a local-first operational workspace that now works with OpenClaw, Hermes, and Codex through MCP.
The idea is to extend natural-language memory with a more explicit operational layer, so agents can persist, inspect, update, and act on structured state over time.
Forge gives agents :
- goals, strategies, projects, issues, tasks, and subtasks
- mixed Kanban + hierarchy views
- live task execution and agent collaboration
- structured notes and wiki memory, with an OpenAI API powered knowledge ingestion
- habits and progress / XP
- calendar and event management
- iCloud / Google Calendar integration
- health-related context like sleep, movement (through iOS companion app in review)
- psychologically relevant structured records
There’s also a Psyche / mental-model layer for things like values, beliefs, patterns, modes, and self-observation reports.
That part is influenced by 3rd-wave CBT, especially ACT and Schema Therapy, but the goal is practical rather than clinical. The point is to help an agent track what matters, what keeps recurring, what creates avoidance or fragmentation, and which actions are aligned with longer-term direction.
In practice, Forge lets Codex can keep an explicit picture of your work and life, connect psychological patterns to concrete projects and actions, and switch between chat-based work and a dedicated UI when visual structure is more useful.
It currently works as:
- an OpenClaw plugin
- a Hermes plugin
- a Codex MCP integration
What I’d most love feedback on:
- what feels genuinely useful vs unnecessary
- any install, setup, or onboarding friction
- whether the Codex surfaces feel coherent
Would love to collaborate with anyone interested.
Website: https://albertbuchard.github.io/forge/
GitHub: https://github.com/albertbuchard/forge
Maintainer: @proven pulsar
civic sorrel