Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors
Rust
fix(wrapper): clarify remote upgrade state
AkitaOnRailsb479d31f711f
diff --git a/CHANGELOG.md b/CHANGELOG.mdindex 68340da1..8e81ba54 100644--- a/CHANGELOG.md+++ b/CHANGELOG.md@@ -7,6 +7,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0## [Unreleased]+### Fixed+- Corrected Docker-wrapper upgrade guidance: `ai-memory upgrade` no longer+ claims a configured remote server is stale when it cannot inspect that+ deployment, and the install guide now makes clear that refreshing Docker+ script hooks does not convert them to native capture-policy commands.+## [1.17.2] - 2026-07-22### Added@@ -29,8 +35,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0the spool/wire. Optional assistant/Stop capture proposed in #196 remainsdisabled. Upgrading the binary is sufficient for native Claude Code installs;script-fallback installs (Docker wrapper, `AI_MEMORY_HOOK_PLATFORM=posix`)- should run `ai-memory install-hooks --agent claude-code --apply` to migrate to- native commands and close the residual local-wire vector ([#196]).+ still send the raw field to the server, which strips it immediately on+ receipt. Closing the local-wire vector requires a native client on the agent+ host and using it to reinstall hooks; running the installer through the+ Docker wrapper only refreshes its scripts ([#196]).### Fixed- Removed the unused `syntect` dependency and its `plist`, `quick-xml`,diff --git a/bin/ai-memory b/bin/ai-memoryindex d22fc875..c960824d 100755--- a/bin/ai-memory+++ b/bin/ai-memory@@ -181,18 +181,16 @@ cmd_upgrade() {fifi- # If the server runs on a different host (homelab scenario), only- # the local wrapper + image + hook scripts got refreshed here.- # Surface that, because the user still needs to deploy on the- # remote box for the server-side binary to update.+ # If the server runs on a different host (homelab scenario), this command+ # cannot know or change that host's deployment state.if [ -n "${AI_MEMORY_SERVER_URL:-}" ] \&& ! echo "${AI_MEMORY_SERVER_URL}" | grep -qE '^https?://(127\.|localhost|\[?::1\]?)'; thenechoecho "→ Note: AI_MEMORY_SERVER_URL points at ${AI_MEMORY_SERVER_URL}"- echo " Your local image is now up to date, but the remote server still"- echo " runs the previous version. Redeploy on that host (e.g. \`bin/deploy\`"- echo " or \`docker compose pull && docker compose up -d\` in its deploy dir)"- echo " for the server-side binary to update."+ echo " ai-memory upgrade updates this wrapper and local image only; it does not"+ echo " inspect or redeploy the remote server. If that host is not already current,"+ echo " run \`bin/deploy\` or \`docker compose pull && docker compose up -d\`"+ echo " in its deploy directory."fimkdir -p "${CACHE_DIR}"diff --git a/crates/ai-memory-cli/tests/packaging.rs b/crates/ai-memory-cli/tests/packaging.rsindex 38d8152d..f977ab60 100644--- a/crates/ai-memory-cli/tests/packaging.rs+++ b/crates/ai-memory-cli/tests/packaging.rs@@ -276,6 +276,44 @@ fn managed_run_wrapper_uses_host_binary_path_and_remote_server_without_docker()assert!(!docker_record.exists(), "managed run entered Docker");}+#[cfg(unix)]+#[test]+fn wrapper_upgrade_does_not_claim_an_updated_remote_server_is_stale() {+ let tmp = tempfile::tempdir().unwrap();+ let docker = tmp.path().join("docker");+ std::fs::write(+ &docker,+ "#!/usr/bin/env bash\n\+ case \"$1\" in\n\+ pull | ps) exit 0 ;;\n\+ *) exit 1 ;;\n\+ esac\n",+ )+ .unwrap();+ {+ use std::os::unix::fs::PermissionsExt as _;+ std::fs::set_permissions(&docker, std::fs::Permissions::from_mode(0o755)).unwrap();+ }++ let output = shell_script_command(&repo_root().join("bin/ai-memory"))+ .arg("upgrade")+ .env("AI_MEMORY_DOCKER", &docker)+ .env("AI_MEMORY_SKIP_SELF_UPGRADE", "1")+ .env("AI_MEMORY_SERVER_URL", "http://192.168.0.90:49374")+ .env("HOME", tmp.path())+ .output()+ .unwrap();+ assert!(+ output.status.success(),+ "wrapper failed: {}",+ String::from_utf8_lossy(&output.stderr)+ );+ let stdout = String::from_utf8_lossy(&output.stdout);+ assert!(stdout.contains("does not\n inspect or redeploy the remote server"));+ assert!(stdout.contains("If that host is not already current"));+ assert!(!stdout.contains("remote server still\n runs the previous version"));+}+#[cfg(unix)]#[test]fn docker_wrapper_completions_tolerate_an_early_reader_close() {diff --git a/docs/install.md b/docs/install.mdindex f1d0db63..27bea4ef 100644--- a/docs/install.md+++ b/docs/install.md@@ -401,8 +401,13 @@ Upgrading the binary is sufficient for native Claude Code installs, and pendingspooled events drain with the field stripped as well. Installs that run the`.sh`/`.ps1` script fallback (the Docker script bundle or an explicit`AI_MEMORY_HOOK_PLATFORM=posix`) still POST the raw field on the local wire-until they move to native commands: run `install-hooks --agent claude-code---apply`, which installs native `ai-memory hook` commands where supported.+until they move to native commands. The Docker wrapper deliberately keeps+script commands because a binary path inside its helper container is not valid+on the host; running `install-hooks` through that wrapper refreshes the scripts+but does not convert them. To close the local-wire exposure, install a native+ai-memory client on the agent host, then use that native executable to run+`install-hooks --agent claude-code --apply`. If the script fallback is retained,+the server still strips the field immediately on receipt before persistence.Native `ai-memory hook --event ...` commands spool events locally. Session startdoes a short bounded cleanup drain before fetching a handoff; cancellation-prone