Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors
Rust
docs(llm): note the Cursor provider's schema-in-prompt and inherited environment
AkitaOnRails00ab34b0d786
diff --git a/docs/llm-providers.md b/docs/llm-providers.mdindex 923b3b73..6824bc96 100644--- a/docs/llm-providers.md+++ b/docs/llm-providers.md@@ -82,7 +82,10 @@ ai-memory llm-test --provider codex --model gpt-5.6-luna --structured --prompt "`AI_MEMORY_CURSOR_AGENT` overrides the binary. The provider writes the promptinto a temporary workspace and calls `agent --print --mode ask --trust`. Itdoes not pass `--yolo` or `--force`: consolidation text is untrusted captured-content, and those flags would allow tool execution.+content, and those flags would allow tool execution. The Cursor CLI has no+JSON-schema mode, so the schema travels in the prompt and the answer is parsed+and validated against it. The `agent` process inherits the server's+environment, so run the server with only the credentials it needs.```bashexport AI_MEMORY_LLM_PROVIDER=cursor