nikira-studio/agent-core

Self-hosted shared memory, encrypted credentials, and scoped connectors for AI agents. MCP and HTTP for Claude Code, Codex, Cursor, and custom runtimes.

0Stars on the repository
3Mods indexed here, across every type
1mo agoLast push, which is what freshness is scored on
MITLicence, which decides whether bodies are shown

nikira-studio/agent-core

Skill Claude CodeCodex

List connector types and bindings before executing an action. Use connectorsrun with a stored binding. Do not request or expose raw secrets. A delegated grant can limit one exact binding and action, and cannot exceed the receiving agent's standing capabilities.

not rated 0 1mo ago A 0 tokens original MIT

nikira-studio/agent-core

Skill Claude CodeCodex

Start with workspacesync for the workspace scope and reuse its execution ID. Search memory for stable project facts and decisions. Write activity updates while working. A fresh activity can return a bounded sincelastactive fallback digest when the execution is not caught up; process it without treating it as complete…

not rated 0 changed 1mo ago A 0 tokens original MIT

agent-core-memory

03

nikira-studio/agent-core

Skill Claude CodeCodex

Search memory before writing. Store durable facts and decisions, not routine progress. Keep facts checkable with a subject anchor. Treat duplicate and previously-retracted write warnings as prompts to inspect, not as failures. Confirm a fact only after checking it, and include evidence. Use activity updates for work…

not rated 0 changed 1mo ago A 0 tokens original MIT