Memory Layer
Agents

Agents

Agents need project context before making assumptions. Memory Layer exposes context through CLI commands, repo-local instructions, watchers, and MCP tools.

Memory Layer agents dashboard

Codex CLI
Claude Code
Generic agent
Agent install prompt
Agent briefings
Permissions and safety
  1. Query project memory before making project-specific claims.
  2. Cite memory IDs or source evidence when using remembered facts.
  3. Verify code and docs before edits.
  4. Store durable discoveries after meaningful work.

Next

Read Codex CLI, Generic agent, and MCP.

Trust and staleness

Handle stale memories with evidence, verification, and replacement proposals.

Codex CLI

Use Memory Layer from Codex workflows with project-local skills and evidence-first prompts.

On this page

AgentsRecommended agent behaviorNext