Repository Issue Activity (beta)

#449 Proposal: optional endpoint-bound preflight hook before remote MCP calls

#2 Dependency Dashboard

#350 Make the policy decision point pluggable (the rules are injectable, the evaluator is not)

#484 A routine's run is governed as though somebody were watching it

#532 The LlamaIndex Bot only answers with an OpenAI model sent to OpenAI

#529 Every allowed action on a dry-run deployment's Audit page says it was not enforced

#459 A deployment with no routines worker is indistinguishable from one running them

#527 A tool argument that starts with "Basic" or "Bearer" is refused as a credential

#322 [Feature Request] Please add Agent templates - import/share

#431 Support Feishu as an external Bot channel

#295 ACP provider Bot: let a coworker run on a coding harness somebody already pays for

#280 [Feature Request] Opt-in self-improving memory for all Agent's

#522 Optional durable company and project memory for OpenBot coworkers

#219 Add built-in agent acceptance test and bounded local Hermes bridge

#349 [Feature Request] API-only LLM access makes OpenBot unnecessarily limited

#357 Publish agent-langgraph alongside openbot on release

#495 DELETE /api/plugins/grants accepts whitespace ref/agentId and reports ok:true without deleting anything

#497 numberFromEnv accepts fractions and out-of-range ports instead of taking the fallback

#496 GET /channels and GET /api/admin/people coerce malformed limit instead of refusing it

#494 POST /api/plugins/skills accepts a non-string slug/summary and answers 500 instead of 400

#465 The People screen forgets when somebody last signed in

#467 `bun run dev` starts a routines worker that throws at boot

#479 /desks registers the live seat twice and CopilotKit throws

#461 An identity provider being registered leaves no audit row, though removing one does

#447 A tool line names its server in raw catalogue form, and the carousel shows two dead arrows with four agents or fewer