Skip to content

Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors

Rust

fix(run): preflight ai-jail availability

AkitaOnRails7f8b95fea545
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 8e0b0aae..6eb549be 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -24,7 +24,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
- Fixed `ai-memory run --jail` leaving the `ssh` toggle unchecked when
`git push` goes over SSH but `origin`'s fetch URL does not show it: an HTTPS
`origin` with an SSH `pushurl` or `pushInsteadOf`, or a `host:path` remote
- using an `~/.ssh/config` alias. (#1024)
+ using an `~/.ssh/config` alias. The interactive offer now also uses
+ ai-jail's own dry-run preflight, so a backend that exists but fails ai-jail's
+ trust checks is treated as unavailable instead of producing a broken offer.
+ (#1024)
## [2.5.2] - 2026-10-01
diff --git a/crates/ai-memory-cli/tests/suite/packaging.rs b/crates/ai-memory-cli/tests/suite/packaging.rs
index 622df167..2101888d 100644
--- a/crates/ai-memory-cli/tests/suite/packaging.rs
+++ b/crates/ai-memory-cli/tests/suite/packaging.rs
@@ -2301,6 +2301,7 @@ mod slow {
.arg("upgrade")
.env("PATH", path)
.env("HOME", tmp.path())
+ .env("XDG_DATA_HOME", tmp.path().join("data"))
.env("AI_MEMORY_DOCKER", &docker)
.env(
"AI_MEMORY_WRAPPER_URL",
@@ -2726,6 +2727,7 @@ mod slow {
.env("AI_MEMORY_SKIP_SELF_UPGRADE", "1")
.env("AI_MEMORY_SERVER_URL", "http://192.168.0.90:49374")
.env("HOME", tmp.path())
+ .env("XDG_DATA_HOME", tmp.path().join("data"))
.output()
.unwrap();
assert!(
diff --git a/crates/ai-memory-workstream/src/jail.rs b/crates/ai-memory-workstream/src/jail.rs
index 3a191251..b4b7cb23 100644
--- a/crates/ai-memory-workstream/src/jail.rs
+++ b/crates/ai-memory-workstream/src/jail.rs
@@ -54,11 +54,35 @@ pub fn usable_ai_jail(os: JailOs, lookup: impl Fn(&str) -> Option<PathBuf>) -> O
/// install location when that directory is not on `PATH`).
#[must_use]
pub fn usable_ai_jail_here() -> Option<PathBuf> {
- usable_ai_jail(current_jail_os(), |name| match find_on_path(name) {
+ let candidate = usable_ai_jail(current_jail_os(), |name| match find_on_path(name) {
Some(path) => Some(path),
+ None if name == "bwrap" => std::env::var_os("BWRAP_BIN").map(PathBuf::from),
None if name == "ai-jail" => home_local_bin_ai_jail(),
None => None,
- })
+ })?;
+ preflighted_ai_jail(candidate, ai_jail_preflight)
+}
+
+fn preflighted_ai_jail(
+ candidate: PathBuf,
+ preflight: impl FnOnce(&Path) -> bool,
+) -> Option<PathBuf> {
+ preflight(&candidate).then_some(candidate)
+}
+
+/// Ask the installed ai-jail to validate its complete local launch boundary.
+/// `--dry-run` does not execute the child or write configuration, but it does
+/// apply ai-jail's own backend trust checks (including root ownership and the
+/// protected-Nix-store exception) so our offer cannot lead directly to a
+/// refusal that the shallower PATH probe missed.
+fn ai_jail_preflight(ai_jail: &Path) -> bool {
+ std::process::Command::new(ai_jail)
+ .args(["--dry-run", "--", "/bin/true"])
+ .stdin(std::process::Stdio::null())
+ .stdout(std::process::Stdio::null())
+ .stderr(std::process::Stdio::null())
+ .status()
+ .is_ok_and(|status| status.success())
}
fn find_on_path(name: &str) -> Option<PathBuf> {
@@ -872,6 +896,22 @@ mod tests {
assert_eq!(usable_ai_jail(JailOs::MacOs, linux_backend_on_macos), None);
}
+ #[test]
+ fn failed_ai_jail_preflight_suppresses_the_offer() {
+ let candidate = PathBuf::from("/opt/bin/ai-jail");
+ assert_eq!(
+ preflighted_ai_jail(candidate.clone(), |path| {
+ assert_eq!(path, candidate);
+ false
+ }),
+ None
+ );
+ assert_eq!(
+ preflighted_ai_jail(candidate.clone(), |_| true),
+ Some(candidate)
+ );
+ }
+
/// ai-jail is unsupported on Windows: even a file named `ai-jail` on PATH
/// (a Git-Bash or WSL shim) must not produce the offer, and the lookup is
/// never consulted.
diff --git a/docs/design-yolo-safety-ai-jail.md b/docs/design-yolo-safety-ai-jail.md
index 7a904acd..342075ca 100644
--- a/docs/design-yolo-safety-ai-jail.md
+++ b/docs/design-yolo-safety-ai-jail.md
@@ -58,9 +58,11 @@ scripts, hooks, and CI keep working unchanged.
The offer appears only when accepting it can actually work
(`usable_ai_jail`): on Linux or macOS, with the ai-jail binary on `PATH`
-(fallback `~/.local/bin/ai-jail`) **and** its sandbox backend on `PATH`
-(`bwrap` on Linux, `sandbox-exec` on macOS). It never appears on Windows, where
-ai-jail is unsupported, even if a file named `ai-jail` happens to be on `PATH`.
+(fallback `~/.local/bin/ai-jail`) **and** a sandbox backend candidate available
+(`bwrap` on Linux, including `BWRAP_BIN`; `sandbox-exec` on macOS). A no-exec
+`ai-jail --dry-run` preflight then applies ai-jail's own backend trust and local
+configuration checks before the offer is shown. It never appears on Windows,
+where ai-jail is unsupported, even if a file named `ai-jail` happens to be on `PATH`.
When ai-jail is not usable there is no question at all — the run proceeds
directly after the §1 warning. The re-exec runs the exact path this check
resolved, never a bare `ai-jail` re-looked-up through `PATH` (which missed a