ur-agent 1.81.0 → 1.81.2
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 +26 -0
- package/dist/cli.js +614 -451
- package/docs/REDTEAM.md +6 -0
- package/docs/VALIDATION.md +1 -1
- package/docs/providers.md +17 -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/plugins/core/reverse-skills/.ur-plugin/plugin.json +1 -1
- package/plugins/core/reverse-skills/README.md +6 -0
- package/plugins/core/reverse-skills/UR-INTEGRATION.md +8 -7
- package/plugins/core/reverse-skills/commands/start.md +10 -0
- package/plugins/core/reverse-skills/skills/application-redteam/SKILL.md +1 -1
- package/plugins/core/reverse-skills/skills/binary-reverse/SKILL.md +1 -1
- package/plugins/core/reverse-skills/skills/exploit-development/SKILL.md +1 -1
- package/plugins/core/reverse-skills/skills/forensics-threat-hunting/SKILL.md +1 -1
- package/plugins/core/reverse-skills/skills/llm-agent-security/SKILL.md +1 -1
- package/plugins/core/reverse-skills/skills/malware-edr-research/SKILL.md +1 -1
- package/plugins/core/reverse-skills/skills/platform-radio-security/SKILL.md +1 -1
- package/plugins/core/reverse-skills/skills/research-evidence/SKILL.md +1 -1
- package/plugins/core/reverse-skills/skills/reverse-skill-router/SKILL.md +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,31 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## 1.81.2
|
|
4
|
+
|
|
5
|
+
- Fixed `/effort max` being acknowledged as `max` while OpenRouter requests,
|
|
6
|
+
the spinner, and the status indicator still used `high`. UR now preserves
|
|
7
|
+
provider reasoning metadata, treats `max` as the model's highest capability,
|
|
8
|
+
and translates it to the correct OpenRouter wire value (including `xhigh`).
|
|
9
|
+
- Made the live provider/model selection authoritative for the next request and
|
|
10
|
+
removed the duplicate `/model` settings write, preventing a newly selected
|
|
11
|
+
OpenRouter model from being dispatched through stale Ollama settings.
|
|
12
|
+
- Bounded deterministic duplicate Edit, Write, and NotebookEdit failures. After
|
|
13
|
+
two unchanged failures UR tells the model to reread the target and rebuild a
|
|
14
|
+
unique edit; continued identical calls stop the turn instead of looping.
|
|
15
|
+
- Replaced the sparse welcome-screen block mark with a balanced outlined UR
|
|
16
|
+
wordmark and branded tagline, while keeping a concise plain-text rendering in
|
|
17
|
+
screen-reader mode.
|
|
18
|
+
|
|
19
|
+
## 1.81.1
|
|
20
|
+
|
|
21
|
+
- Preflighted UR task tracking in `/reverse-skills:start` and every bundled
|
|
22
|
+
reverse-security skill before mutating, multi-outcome workflows. The plugin
|
|
23
|
+
now creates and advances an actionable task board instead of first tripping
|
|
24
|
+
`TaskListRequired`, while keeping `tasks.requireBeforeChanges` enabled.
|
|
25
|
+
- Made verifier and command-registry tests independent of globally installed
|
|
26
|
+
plugins, so release checks remain reproducible when marketplace validators
|
|
27
|
+
and namespaced commands are enabled on the developer machine.
|
|
28
|
+
|
|
3
29
|
## 1.81.0
|
|
4
30
|
|
|
5
31
|
- Added `/mode redteam`, a warning-gated, session-only UR policy mode for
|