feat(uninstall): add oo uninstall command#245
Merged
Conversation
Introduce `oo uninstall` to remove the managed oo runtime and its oo-owned skills, mirroring what `oo install` puts in place. By default it removes the managed executable, installed versions, self-update staging and locks, bundled skills, and the preset registry skill package, while leaving PATH config, user-authored skills, and unmanaged same-name directories untouched. Skill removal is gated on `.oo-metadata.json` ownership so user-authored same-name skills are never deleted. `--purge` also clears user data and all oo-managed registry skills, `--dry-run` prints the plan without deleting, and `-y/--yes` is required in non-interactive terminals. On Windows the running executable is removed via a deferred one-shot PowerShell helper since a process cannot delete itself in place. The preset skill package list moves into a shared preset-packages module so install and uninstall share a single source of truth, and the command refuses to run while another live oo process holds an active version marker. Signed-off-by: Kevin Cui <bh@bugs.cc>
Summary by CodeRabbit
WalkthroughThis PR implements a complete Sequence Diagram(s)sequenceDiagram
participant User
participant CLI as oo uninstall
participant Plan as buildSelfUninstallPlan
participant Lock as findAnyActiveVersionOwner
participant Perform as performSelfUninstall
participant Helper as Windows Cleanup
participant Filesystem as File System
User->>CLI: oo uninstall --yes --purge
CLI->>Plan: analyze skills/runtime/data
Plan->>Plan: classify by metadata
Plan-->>CLI: UninstallPlan
CLI->>Perform: execute plan
Perform->>Lock: check active owner
alt owner found
Lock-->>Perform: ownerPid
Perform-->>CLI: error (busy)
else no owner
Lock-->>Perform: undefined
alt Windows + deferred items
Perform->>Helper: spawn PowerShell script
Helper->>Filesystem: wait for process
Helper->>Filesystem: delete deferred paths
Helper->>Filesystem: self-delete
end
Perform->>Filesystem: rm immediate items
Filesystem-->>Perform: failedPaths[]
Perform-->>CLI: result
end
CLI-->>User: success or error
Possibly related PRs
🚥 Pre-merge checks | ✅ 4✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches✨ Simplify code
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (5)
docs/commands.md (1)
433-440: ⚡ Quick winKeep uninstall docs at user-facing behavior level.
This section exposes internal mechanics (“one-shot PowerShell helper” and “active version marker”). Please rephrase to externally observable behavior (deferred Windows cleanup after process exit; refuses when another
ooprocess is running).As per coding guidelines, “Documentation under
docs/commands*.mdshould describe only user-facing CLI contract ... Do not document internal implementation details.”🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/commands.md` around lines 433 - 440, Reword the uninstall docs to remove internal implementation details: replace the Windows sentence that mentions "`~/.local/bin/oo.exe`" and "one-shot PowerShell helper" with a user-facing statement like "On Windows the executable is removed after the process exits (cleanup deferred)," and replace the sentence mentioning "active version marker" with a user-facing statement that the command "refuses to run while another live `oo` process is running"; do not mention helper scripts, file paths, markers, or other internal mechanics and keep the description focused on observable CLI behavior.docs/commands.zh-CN.md (1)
371-376: ⚡ Quick win文档建议改为仅描述用户可见行为。
这里写到了内部实现细节（一次性 PowerShell helper、active version marker）。建议改成用户可观察到的结果描述（Windows 下进程退出后再删除可执行文件；检测到其他
oo进程运行时拒绝执行）。As per coding guidelines, “Documentation under
docs/commands*.mdshould describe only user-facing CLI contract ... Do not document internal implementation details.”🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/commands.zh-CN.md` around lines 371 - 376, 将文档从描述内部实现细节（例如“一次性 PowerShell helper”、“active version marker”）改为只描述用户可见行为：在涉及 `~/.local/bin/oo.exe` 的段落，移除对 PowerShell helper 的说明，改写为“Windows 下可执行文件将在进程退出后被删除；其它路径（versions、staging、locks、skills）会在命令执行期间被删除/清理”；在安全段落，移除对 `active version marker` 的内部名词，改写为“当检测到另一个运行中的 oo 进程时，命令会拒绝执行”；保留关于不将 API key 等机密写入 stdout/stderr 的用户可观察承诺。确保只描述可见行为并删除任何实现细节或内部术语。src/application/commands/uninstall.cli.test.ts (3)
79-307: ⚡ Quick winExtract the repeated sandbox setup/cleanup pattern into a local factory helper.
Each test repeats the same sandbox lifecycle; a small local helper will reduce duplication and maintenance churn.
As per coding guidelines, "In test files, extract repeated setup (mock, stub, or setup objects) into a local factory function at the bottom of the file."
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/application/commands/uninstall.cli.test.ts` around lines 79 - 307, Tests repeat the same sandbox lifecycle (createCliSandbox(), try { ... await sandbox.cleanup(); } finally) — introduce a local factory helper (e.g., runWithSandbox or withCliSandbox) at the bottom of this test file that creates the sandbox, yields it to an async callback, and guarantees cleanup in a finally block; update each test to call this helper and move usages of createCliSandbox, sandbox.run, seedRuntime, seedHostSkill, and sandbox.cleanup into the callback so the helper centralizes setup/teardown and reduces duplication.
21-75: ⚡ Quick winMove file-local test helpers to the end of the test file.
These helpers appear file-scoped; placing them at the bottom keeps test flow first and aligns with repo test conventions.
As per coding guidelines, "If a test helper function might be called by other test files, place it in the
__tests__/helpers.tsfile. Otherwise, place the function in the test file at the end of that file."🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/application/commands/uninstall.cli.test.ts` around lines 21 - 75, Move the file-local test helper functions selfUpdatePaths, storePaths, seedRuntime, seedHostSkill, and nativeExecPath to the bottom of this test file (after all tests) so tests appear first; if any of these helpers are referenced by other test files, instead extract them into __tests__/helpers.ts and import them from there, keeping their signatures unchanged so calls in this file still resolve.
19-19: ⚡ Quick winReuse the shared preset package source instead of hardcoding the package string.
Keeping this value local can drift from the shared preset package list used by install/uninstall logic.
As per coding guidelines, "Never duplicate constant values across files. Define once, import or re-export with aliases elsewhere."
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/application/commands/uninstall.cli.test.ts` at line 19, Replace the hardcoded PRESET_PACKAGE constant in uninstall.cli.test.ts with the shared preset package constant used by the install/uninstall logic: remove the local const PRESET_PACKAGE = "`@alwaysmavs/gpt-image-2`" and import the canonical preset identifier exported by the shared presets module (the same symbol used by install/uninstall code) so the test reuses the single source of truth for preset package names; update references to use that imported symbol (e.g., PRESET_PACKAGE or PRESetsMap key) to avoid duplication and drift.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@src/application/self-update/lock.ts`:
- Around line 229-238: The loop over versionEntries should skip non-directory
entries and not abort on a stray file; before calling
findActiveMarkerOwner(join(activeDirectory, versionEntry), options) use fs.lstat
or fs.stat to verify the path is a directory and if not continue, and also guard
the call with a try/catch that logs/ignores errors and continues the loop so a
thrown error from findActiveMarkerOwner doesn't break the scan; update the loop
that references versionEntries, activeDirectory, and findActiveMarkerOwner
accordingly.
In `@src/application/self-update/uninstall.test.ts`:
- Line 5: Replace hardcoded Windows path separators in the test assertions with
path helpers: import and use join(...) (already imported) to construct expected
paths (or use path.relative(...) to compare relationships) instead of asserting
raw strings with backslashes; update the assertions in uninstall.test.ts (the
Windows-path expectations around the self-update/uninstall tests) to build
expected paths via join() so they become OS-neutral and resilient.
---
Nitpick comments:
In `@docs/commands.md`:
- Around line 433-440: Reword the uninstall docs to remove internal
implementation details: replace the Windows sentence that mentions
"`~/.local/bin/oo.exe`" and "one-shot PowerShell helper" with a user-facing
statement like "On Windows the executable is removed after the process exits
(cleanup deferred)," and replace the sentence mentioning "active version marker"
with a user-facing statement that the command "refuses to run while another live
`oo` process is running"; do not mention helper scripts, file paths, markers, or
other internal mechanics and keep the description focused on observable CLI
behavior.
In `@docs/commands.zh-CN.md`:
- Around line 371-376: 将文档从描述内部实现细节（例如“一次性 PowerShell helper”、“active version
marker”）改为只描述用户可见行为：在涉及 `~/.local/bin/oo.exe` 的段落，移除对 PowerShell helper
的说明，改写为“Windows
下可执行文件将在进程退出后被删除；其它路径（versions、staging、locks、skills）会在命令执行期间被删除/清理”；在安全段落，移除对
`active version marker` 的内部名词，改写为“当检测到另一个运行中的 oo 进程时，命令会拒绝执行”；保留关于不将 API key
等机密写入 stdout/stderr 的用户可观察承诺。确保只描述可见行为并删除任何实现细节或内部术语。
In `@src/application/commands/uninstall.cli.test.ts`:
- Around line 79-307: Tests repeat the same sandbox lifecycle
(createCliSandbox(), try { ... await sandbox.cleanup(); } finally) — introduce a
local factory helper (e.g., runWithSandbox or withCliSandbox) at the bottom of
this test file that creates the sandbox, yields it to an async callback, and
guarantees cleanup in a finally block; update each test to call this helper and
move usages of createCliSandbox, sandbox.run, seedRuntime, seedHostSkill, and
sandbox.cleanup into the callback so the helper centralizes setup/teardown and
reduces duplication.
- Around line 21-75: Move the file-local test helper functions selfUpdatePaths,
storePaths, seedRuntime, seedHostSkill, and nativeExecPath to the bottom of this
test file (after all tests) so tests appear first; if any of these helpers are
referenced by other test files, instead extract them into __tests__/helpers.ts
and import them from there, keeping their signatures unchanged so calls in this
file still resolve.
- Line 19: Replace the hardcoded PRESET_PACKAGE constant in
uninstall.cli.test.ts with the shared preset package constant used by the
install/uninstall logic: remove the local const PRESET_PACKAGE =
"`@alwaysmavs/gpt-image-2`" and import the canonical preset identifier exported by
the shared presets module (the same symbol used by install/uninstall code) so
the test reuses the single source of truth for preset package names; update
references to use that imported symbol (e.g., PRESET_PACKAGE or PRESetsMap key)
to avoid duplication and drift.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: ea4be4b9-9f3e-414c-81c5-7f6eda521181
⛔ Files ignored due to path filters (2)
src/application/bootstrap/__snapshots__/run-cli.test.ts.snapis excluded by!**/*.snapsrc/application/commands/config/__snapshots__/index.cli.test.ts.snapis excluded by!**/*.snap
📒 Files selected for processing (13)
docs/commands.mddocs/commands.zh-CN.mdsrc/application/commands/catalog.tssrc/application/commands/skills/index.test.tssrc/application/commands/skills/install.tssrc/application/commands/skills/preset-packages.tssrc/application/commands/telemetry-decisions.test.tssrc/application/commands/uninstall.cli.test.tssrc/application/commands/uninstall.tssrc/application/self-update/lock.tssrc/application/self-update/uninstall.test.tssrc/application/self-update/uninstall.tssrc/i18n/catalog.ts
Comment on lines
+229
to
+238
| for (const versionEntry of versionEntries) { | ||
| const owner = await findActiveMarkerOwner( | ||
| join(activeDirectory, versionEntry), | ||
| options, | ||
| ); | ||
|
|
||
| if (owner !== undefined) { | ||
| return owner; | ||
| } | ||
| } |
There was a problem hiding this comment.
Ignore non-directory entries while scanning active lock versions.
A stray file under locks/active can make findActiveMarkerOwner throw, which aborts the busy-check instead of continuing scan.
💡 Proposed hardening
for (const versionEntry of versionEntries) {
- const owner = await findActiveMarkerOwner(
- join(activeDirectory, versionEntry),
- options,
- );
+ let owner: { ownerPid: number } | undefined;
+ try {
+ owner = await findActiveMarkerOwner(
+ join(activeDirectory, versionEntry),
+ options,
+ );
+ }
+ catch (error) {
+ if (isDirectoryReadError(error) || isPathMissingError(error)) {
+ continue;
+ }
+ throw error;
+ }
if (owner !== undefined) {
return owner;
}
}🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@src/application/self-update/lock.ts` around lines 229 - 238, The loop over
versionEntries should skip non-directory entries and not abort on a stray file;
before calling findActiveMarkerOwner(join(activeDirectory, versionEntry),
options) use fs.lstat or fs.stat to verify the path is a directory and if not
continue, and also guard the call with a try/catch that logs/ignores errors and
continues the loop so a thrown error from findActiveMarkerOwner doesn't break
the scan; update the loop that references versionEntries, activeDirectory, and
findActiveMarkerOwner accordingly.
On Windows, `oo uninstall --purge` exited non-zero because the running process holds the SQLite databases under the data directory open, and Windows refuses to delete files held open by a live process. The data directory is now handed to the post-exit PowerShell helper alongside the running executable, so `--purge` completes cleanly; config files, logs, and telemetry are still removed in-process. The helper directory is now resolved for every Windows plan so deferred paths have a post-exit home. This also fixes the Windows CI test suite: - the CLI test now expects the deferred "scheduled" message and a retained executable on Windows, where the running image cannot self-delete; - buildSelfUninstallPlan tests build the exec path with the simulated platform's separators so native detection matches on a Windows runner; - the failedPaths test uses an embedded-NUL path that fs.rm rejects on every platform instead of the ENOTDIR trick Windows tolerates. Docs updated to describe the broadened Windows deferral. Signed-off-by: Kevin Cui <bh@bugs.cc>
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/application/self-update/uninstall.ts (1)
371-386:⚠️ Potential issue | 🟠 Major | ⚡ Quick winKeep the active-owner guard independent from the deletion plan.
Because
addRuntimeItemsonly runs for native installs on Line 136, non-native plans never contain a"locks"item. That makeslocksItemundefined here and silently skips the busy-process check, even though uninstall is supposed to refuse while another liveooprocess still owns an active-version marker. This can let a package-manager install purge shared skills/user data while anotherooprocess is running.Suggested fix
export interface UninstallPlan { + activeVersionLocksDirectory: string; /** * Items whose removal must be deferred to a post-exit helper (the Windows * running image cannot unlink itself in-process). Membership in this array, * not a per-item flag, is what marks an item as deferred. */ deferred: UninstallPlanItem[]; @@ return { + activeVersionLocksDirectory: selfUpdatePaths.locksDirectory, deferred, helperDirectory, immediate, installationMethod, @@ export async function performSelfUninstall( options: PerformUninstallOptions, ): Promise<PerformUninstallResult> { - const locksItem = options.plan.immediate.find(item => item.category === "locks"); - - if (locksItem !== undefined) { - const owner = await findAnyActiveVersionOwner({ - excludeProcessId: options.processId, - locksDirectory: locksItem.path, - platform: options.plan.platform, - }); - - if (owner !== undefined) { - return { - ownerPid: owner.ownerPid, - status: "busy", - }; - } + const owner = await findAnyActiveVersionOwner({ + excludeProcessId: options.processId, + locksDirectory: options.plan.activeVersionLocksDirectory, + platform: options.plan.platform, + }); + + if (owner !== undefined) { + return { + ownerPid: owner.ownerPid, + status: "busy", + }; }🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/application/self-update/uninstall.ts` around lines 371 - 386, The busy-process check must not depend on options.plan.immediate containing a "locks" item; update uninstall.ts so the active-owner guard always runs: call findAnyActiveVersionOwner (using options.processId and a proper locksDirectory derived independently of options.plan.immediate — e.g. the runtime locks path used by addRuntimeItems or a default locks path for the platform) and return { ownerPid, status: "busy" } if it finds an owner. In short, move/replicate the findAnyActiveVersionOwner invocation out of the locksItem conditional and ensure you provide a valid locksDirectory when locksItem is undefined.
🧹 Nitpick comments (1)
src/application/self-update/uninstall.ts (1)
173-176: 💤 Low valueDrop the pass-through
pathsalias.
const paths = args.paths;does not add meaning here; usingargs.pathsdirectly keeps this helper aligned with the repo rule.As per coding guidelines, "Don't create variables that pass through a value unchanged. If no transformation occurs, use the original directly."
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/application/self-update/uninstall.ts` around lines 173 - 176, Remove the unnecessary pass-through alias by deleting the line "const paths = args.paths;" and update any subsequent references to the local variable "paths" to use "args.paths" directly (this affects the function that receives the parameter object and any code inside that uses the "paths" identifier). Ensure no other transformations are needed and that TypeScript types remain satisfied for the parameter "paths: SelfUpdatePaths".
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In `@src/application/self-update/uninstall.ts`:
- Around line 371-386: The busy-process check must not depend on
options.plan.immediate containing a "locks" item; update uninstall.ts so the
active-owner guard always runs: call findAnyActiveVersionOwner (using
options.processId and a proper locksDirectory derived independently of
options.plan.immediate — e.g. the runtime locks path used by addRuntimeItems or
a default locks path for the platform) and return { ownerPid, status: "busy" }
if it finds an owner. In short, move/replicate the findAnyActiveVersionOwner
invocation out of the locksItem conditional and ensure you provide a valid
locksDirectory when locksItem is undefined.
---
Nitpick comments:
In `@src/application/self-update/uninstall.ts`:
- Around line 173-176: Remove the unnecessary pass-through alias by deleting the
line "const paths = args.paths;" and update any subsequent references to the
local variable "paths" to use "args.paths" directly (this affects the function
that receives the parameter object and any code inside that uses the "paths"
identifier). Ensure no other transformations are needed and that TypeScript
types remain satisfied for the parameter "paths: SelfUpdatePaths".
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 4589e276-f997-434a-b496-ad3f607fae08
📒 Files selected for processing (5)
docs/commands.mddocs/commands.zh-CN.mdsrc/application/commands/uninstall.cli.test.tssrc/application/self-update/uninstall.test.tssrc/application/self-update/uninstall.ts
✅ Files skipped from review due to trivial changes (1)
- docs/commands.zh-CN.md
🚧 Files skipped from review as they are similar to previous changes (3)
- src/application/self-update/uninstall.test.ts
- docs/commands.md
- src/application/commands/uninstall.cli.test.ts
The active-version owner safety check was previously gated on the uninstall plan containing a "locks" removal item. When the plan omitted that item, the guard was silently skipped, letting uninstall proceed while another live `oo` process was still running. Store the locks directory on the plan so the owner check always runs regardless of which paths the plan removes. Docs now describe the user-facing contract only, and tests no longer hardcode preset skill package names or non-portable Windows paths. Signed-off-by: Kevin Cui <bh@bugs.cc>
There was a problem hiding this comment.
🧹 Nitpick comments (1)
src/application/self-update/uninstall.test.ts (1)
44-59: ⚡ Quick winMove test helper functions to the end of the file.
The helper functions
categoriesOf,paths, andwriteSkillshould be relocated to the end of the file (after the last test, nearwriteJsonFileat line 477). As per coding guidelines, test-local helper functions belong at the bottom to keep test logic prominent and implementation details secondary.♻️ Suggested relocation
Move these three functions to appear just before or after
writeJsonFileat line 477. ThewriteJsonFilehelper is already correctly placed at the end.-function categoriesOf(items: readonly UninstallPlanItem[]): string[] { - return items.map(item => item.category); -} - -function paths(items: readonly UninstallPlanItem[]): string[] { - return items.map(item => item.path); -} - -async function writeSkill( - skillDirectory: string, - metadataJson: string, -): Promise<void> { - await mkdir(skillDirectory, { recursive: true }); - await writeFile(join(skillDirectory, "SKILL.md"), "# x\n"); - await writeFile(resolveManagedSkillMetadataFilePath(skillDirectory), metadataJson); -} - describe("shouldRemoveManagedSkill", () => {Then add them at the end before or after
writeJsonFile.As per coding guidelines,
**/*.test.{ts,tsx,js,jsx}: If a test helper function might be called by other test files, place it in the__tests__/helpers.tsfile. Otherwise, place the function in the test file at the end.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/application/self-update/uninstall.test.ts` around lines 44 - 59, The helper functions categoriesOf, paths, and writeSkill are declared mid-file and should be moved to the end of the test file to keep tests first; cut the functions named categoriesOf, paths, and writeSkill from their current location and paste them after the other test-local helpers near writeJsonFile (i.e., at the bottom of the file, either just before or after writeJsonFile), preserving their implementations and imports so other tests in this file still call them correctly.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In `@src/application/self-update/uninstall.test.ts`:
- Around line 44-59: The helper functions categoriesOf, paths, and writeSkill
are declared mid-file and should be moved to the end of the test file to keep
tests first; cut the functions named categoriesOf, paths, and writeSkill from
their current location and paste them after the other test-local helpers near
writeJsonFile (i.e., at the bottom of the file, either just before or after
writeJsonFile), preserving their implementations and imports so other tests in
this file still call them correctly.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 997de1ff-814d-46ea-a361-6c48c2420647
📒 Files selected for processing (5)
docs/commands.mddocs/commands.zh-CN.mdsrc/application/commands/uninstall.cli.test.tssrc/application/self-update/uninstall.test.tssrc/application/self-update/uninstall.ts
✅ Files skipped from review due to trivial changes (2)
- docs/commands.md
- docs/commands.zh-CN.md
🚧 Files skipped from review as they are similar to previous changes (2)
- src/application/commands/uninstall.cli.test.ts
- src/application/self-update/uninstall.ts
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Introduce
oo uninstallto remove the managed oo runtime and its oo-owned skills, mirroring whatoo installputs in place. By default it removes the managed executable, installed versions, self-update staging and locks, bundled skills, and the preset registry skill package, while leaving PATH config, user-authored skills, and unmanaged same-name directories untouched.Skill removal is gated on
.oo-metadata.jsonownership so user-authored same-name skills are never deleted.--purgealso clears user data and all oo-managed registry skills,--dry-runprints the plan without deleting, and-y/--yesis required in non-interactive terminals. On Windows the running executable is removed via a deferred one-shot PowerShell helper since a process cannot delete itself in place.The preset skill package list moves into a shared preset-packages module so install and uninstall share a single source of truth, and the command refuses to run while another live oo process holds an active version marker.