ur-agent 1.60.0 → 1.61.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +34 -0
- package/dist/cli.js +1144 -857
- package/docs/USAGE.md +4 -0
- package/documentation/index.html +1 -1
- package/extensions/jetbrains-ur/build.gradle.kts +1 -1
- package/extensions/vscode-ur-inline-diffs/package.json +1 -1
- package/package.json +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,39 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## 1.61.0
|
|
4
|
+
|
|
5
|
+
- Added `ur selftest`, end-to-end drills for the gap that produced every
|
|
6
|
+
serious defect in recent releases: a module that is correct while something
|
|
7
|
+
between it and the user is not — a wire format, a CLI registration, an exit
|
|
8
|
+
code, a wording choice. The drills deliberately do not import the modules;
|
|
9
|
+
they spawn the shipped binary against real directories and assert on what a
|
|
10
|
+
user would see.
|
|
11
|
+
- Five drills run automatically and cover the four defects that reached you:
|
|
12
|
+
memory integrity reporting an unchecked store as verified, the trajectory
|
|
13
|
+
gate printing FAILED while exiting 0, `--costs` hiding which directory it
|
|
14
|
+
searched, and per-agent attribution itself.
|
|
15
|
+
- Drills that need a live model cannot be automated, so they are emitted as
|
|
16
|
+
prompts with the specific observation that separates working from broken —
|
|
17
|
+
"it seemed fine" is how the screenshot bug survived its first report.
|
|
18
|
+
- `ur selftest run` exits non-zero on failure, and a test asserts the drills
|
|
19
|
+
actually fail when pointed at a broken binary. A self-test that cannot fail
|
|
20
|
+
is decoration.
|
|
21
|
+
|
|
22
|
+
## 1.60.1
|
|
23
|
+
|
|
24
|
+
- `ur memory-integrity verify` no longer reports an empty or missing store as
|
|
25
|
+
verified. It printed "verified — 0 file(s) match the recorded digests" for an
|
|
26
|
+
empty directory, and that identical reassurance would have appeared for a
|
|
27
|
+
mistyped path or a store an attacker had just emptied. Zero files checked is
|
|
28
|
+
not evidence of integrity, so empty and nonexistent stores now say exactly
|
|
29
|
+
that and name the path.
|
|
30
|
+
- The exit code tracks tampering rather than validity, so an empty or
|
|
31
|
+
unbaselined store still exits 0. Failing on it would fire for every fresh
|
|
32
|
+
install and train the user to ignore the exit code.
|
|
33
|
+
- `ur agent-inspect --costs` names the directory it searched. "No subagent
|
|
34
|
+
transcripts found for this session" was indistinguishable from having
|
|
35
|
+
resolved the wrong path.
|
|
36
|
+
|
|
3
37
|
## 1.60.0
|
|
4
38
|
|
|
5
39
|
- Added `ur memory-integrity`, tamper-evidence for the file-backed memory
|