kage
01Plugin Claude Code
Kage: a framework for collaborative agent memory.
Plugin Claude Code
Kage: a framework for collaborative agent memory.
Instructions file CodexOpenCode
AGENTS.md instructions for kage-core/Kage, covering kage memory harness, automatic recall, show the value, automatic capture and end-of-task proposal.
Instructions file
Claude Code instructions for kage-core/Kage, covering kage memory harness, automatic recall, show the value, automatic capture and end-of-task proposal.
Plugin Claude Code
Kage manages your memory and agents. It briefs and runs agents in isolated git worktrees for a single run or a multi-wave goal, re-runs the checks itself instead of trusting the agent's report, and remembers what it learns so each brief is smarter than the last.
Command
Show what Kage saved you — tokens, cost, stale memories blocked, stale knowledge caught.
Command
Initialize Kage verified memory in this repo (.agentmemory + code graph).
Command
Run the Kage Truth Report on this repo — duplicates, ghost exports, bus-factor-1 files, knowledge voids, doc lies.
Hook
Runs when a session starts, executing session-start.sh via bash. From kage-core/Kage.
Hook
Runs when you submit a prompt, before the agent sees it, executing observe.sh via bash. From kage-core/Kage.
Hook
Runs before the agent uses a tool for Read, Edit, Write and MultiEdit tool calls, executing observe.sh, kage-read-context.sh and kage-edit-context.sh via bash (3 commands). From kage-core/Kage.
Hook
Runs after a tool call finishes, executing observe.sh via bash. From kage-core/Kage.
Hook
Runs when a tool call fails, executing observe.sh via bash. From kage-core/Kage.
Hook
Runs before the context is compacted, executing observe.sh via bash. From kage-core/Kage.
Hook
Runs when the agent finishes a response, executing stop.sh via bash. From kage-core/Kage.
Hook
Runs when a session ends, executing observe.sh via bash. From kage-core/Kage.
Hook
Runs when a subagent finishes, executing observe.sh via bash. From kage-core/Kage.
Skill Claude CodeCodex
Trustworthy, code-grounded repo memory for coding agents. Before any task, recall grounded repo knowledge (decisions, bugs, conventions, code paths); after learning something durable, capture it; never act on stale or hallucinated memory. Use at the start of every task, when debugging, and after solving something…
MCP server Claude CodeCodexCursor +2
Verified memory for coding agents: claims cited against code, stale withheld, savings receipts. Runs locally from the @kage-core/kage-graph-mcp npm package.