mirror of
https://github.com/runkids/skillshare.git
synced 2026-11-18 18:04:17 +00:00
chore(main): release 0.30.7 (#871)
* chore(main): release 0.30.7 * chore(release): synchronize release metadata * docs: rewrite the 0.30.7 changelog for users Release Please lists commit subjects. The entry now says what each change means for users, with the --respect-accepted example, migration details for install --json, and credit for the reports and fixes. --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Willie <egg358358@gmail.com>
This commit is contained in:
co-authored by
Willie
parent
51089307f8
commit
d04d04fd93
@@ -1,3 +1,3 @@
|
||||
{
|
||||
".": "0.30.6"
|
||||
".": "0.30.7"
|
||||
}
|
||||
|
||||
@@ -1,5 +1,26 @@
|
||||
# Changelog
|
||||
|
||||
## [0.30.7] - 2026-11-13
|
||||
|
||||
### New Features
|
||||
|
||||
- **Let accepted findings pass `skillshare audit`** — findings you accept with `--force` during `install` or `update` are recorded, but `skillshare audit` kept failing on them, so a skill installed that way failed every CI audit. `audit --respect-accepted` applies those acceptances the way the update gate does: the findings stay in the report, marked as accepted, but no longer count toward the exit code. A changed finding needs a new acceptance, and without the flag `audit` behaves as before. JSON marks them `"acknowledged": true` and counts them in `summary.accepted`, Markdown adds `(accepted earlier)`, and SARIF lists them as external suppressions. The dashboard always applies acceptances: the Audit page mutes accepted findings, counts them separately and shows them with **Show accepted**. Acceptances recorded for a tracked repository now also apply to the skills inside it, in `audit` and in the install, update and edit gates, except bundled hook and settings findings. Thanks to @vieGPT for the report. Refs: #869, #872.
|
||||
```bash
|
||||
skillshare audit --respect-accepted
|
||||
skillshare audit --respect-accepted --threshold high --format json
|
||||
```
|
||||
|
||||
### Bug Fixes
|
||||
|
||||
- **Source credentials stay out of the operation log** — `install` logged the source exactly as typed, so a URL such as `https://user:token@host/repo.git` put the token in `operations.log`. Sources are now logged without user info, query or fragment, including where a message repeats them. Thanks to @salmonumbrella for the report. Refs: #878, #881.
|
||||
- **A broken `audit-rules.yaml` no longer turns the audit off** — when the global `audit-rules.yaml` did not parse or one of its patterns did not compile, global scans dropped the error and ran without any pattern rules, so audits silently reported fewer findings. Every scan that uses the rules now reports a scan error that names the file. An install whose audit cannot run is blocked and what it installed is removed, for skills, agents and tracked repositories; `--force` does not override this, and `--skip-audit` still skips scanning. `update` and `upgrade --skill` keep the installed copy, and `update --all` lists the item under "Blocked by security audit" and exits non-zero. Thanks to @salmonumbrella for the report. Refs: #879, #882.
|
||||
- **`install --json` no longer overwrites without `--force`** — `--json` also turned on `--force`, so a script that only wanted JSON output replaced locally edited skills and agents. JSON now changes only the output: existing items are kept and listed in a new `skipped` array, and overwriting needs `--force`. `install -p --json` no longer follows a changed lock pin unless `--force` or `--update` is given. Thanks to @salmonumbrella for the report and the fix. Refs: #877, #884.
|
||||
- **Reinstalling from the same repo is reported as skipped** — the docs say a reinstall from the same repository is skipped with a warning, but the check read only the per-skill file that installs no longer write, so the reinstall failed with "already exists". It now reads `.metadata.json`: reinstalling the same skill of a repo without `--force` exits 0 and lists it under `skipped`. Another skill of that repo that would take the same folder, for example through `--name`, is still a conflict. Reinstalling a single-skill repo with `-s <skill> -a <agent>` still installs the requested agent. In the dashboard, batch installs report kept skills and agents as skipped instead of installed, the install dialog says they are already installed, and a tracked reinstall of the same repo returns 409 instead of an error. The Herdr panel reports skipped names the way it reports failed ones. Refs: #887.
|
||||
- **`skillshare log` shows what an install skipped** — install entries now record skipped skills and agents, shown as `skipped=` in `skillshare log`. This covers CLI installs, `skillshare search` installs, and installs from the dashboard, including config installs. Refs: #886, #887.
|
||||
- **OMP plugins installed into a registered marketplace update again** — installing an Oh My Pi plugin from a marketplace that was already registered used OMP's cached catalog, so OMP recorded an older version and every later update was blocked, while `plugin update --all` skipped it and exited 0. Install now refreshes the marketplace catalog first, and removal follows the version OMP actually recorded once its scope and cache path check out. A target that cannot update is now reported as failed: the other targets still update, `--dry-run` shows the failed rows and exits non-zero, and the plan JSON marks it `failed: true`. Thanks to @salmonumbrella for the report and the fix. Refs: #851, #853.
|
||||
- **Imported plugins no longer fail `plugin update --all`** — the native-support check ran before the rules that intentionally skip a plugin, so an imported Grok plugin made every `plugin update --all` exit non-zero. Intentional skips now stay skips and point to the native client; changes that run a native command are still checked. Refs: #873.
|
||||
- **The Doctor page shows checks as they finish** — the dashboard's Doctor page waited for every check, up to about 30 seconds when a plugin CLI did not answer, and the request was sometimes cut off. Checks now appear as each one finishes, and **Run again** keeps the previous result until the new one is done. The audit, diff and doctor views show the server's reason when a stream fails. API clients can use the new `GET /api/doctor/stream`. Refs: #870.
|
||||
|
||||
## [0.30.6] - 2026-11-11
|
||||
|
||||
### New Features
|
||||
|
||||
@@ -1,3 +1,3 @@
|
||||
// The skillshare release this panel shipped with. Release Please rewrites it in every release
|
||||
// (.github/release-please-config.json), so the panel can tell when a newer release has one.
|
||||
export const VERSION = "0.30.6"; // x-release-please-version
|
||||
export const VERSION = "0.30.7"; // x-release-please-version
|
||||
|
||||
@@ -1,3 +1,3 @@
|
||||
// The skillshare release this panel shipped with. Release Please rewrites it in every release
|
||||
// (.github/release-please-config.json), so the panel can tell when a newer release has one.
|
||||
export const VERSION = "0.30.6"; // x-release-please-version
|
||||
export const VERSION = "0.30.7"; // x-release-please-version
|
||||
|
||||
@@ -9,7 +9,7 @@ description: |
|
||||
noninteractive automation, and guidance for the terminal UI.
|
||||
argument-hint: "[command] [target] [--json] [--dry-run] [-p|-g]"
|
||||
metadata:
|
||||
version: v0.30.6
|
||||
version: v0.30.7
|
||||
---
|
||||
|
||||
# Skillshare CLI
|
||||
|
||||
@@ -9,6 +9,27 @@ All notable changes to skillshare are documented here. For the full commit histo
|
||||
|
||||
---
|
||||
|
||||
## [0.30.7] - 2026-11-13
|
||||
|
||||
### New Features
|
||||
|
||||
- **Let accepted findings pass `skillshare audit`** — findings you accept with `--force` during `install` or `update` are recorded, but `skillshare audit` kept failing on them, so a skill installed that way failed every CI audit. `audit --respect-accepted` applies those acceptances the way the update gate does: the findings stay in the report, marked as accepted, but no longer count toward the exit code. A changed finding needs a new acceptance, and without the flag `audit` behaves as before. JSON marks them `"acknowledged": true` and counts them in `summary.accepted`, Markdown adds `(accepted earlier)`, and SARIF lists them as external suppressions. The dashboard always applies acceptances: the Audit page mutes accepted findings, counts them separately and shows them with **Show accepted**. Acceptances recorded for a tracked repository now also apply to the skills inside it, in `audit` and in the install, update and edit gates, except bundled hook and settings findings. Thanks to @vieGPT for the report. Refs: #869, #872.
|
||||
```bash
|
||||
skillshare audit --respect-accepted
|
||||
skillshare audit --respect-accepted --threshold high --format json
|
||||
```
|
||||
|
||||
### Bug Fixes
|
||||
|
||||
- **Source credentials stay out of the operation log** — `install` logged the source exactly as typed, so a URL such as `https://user:token@host/repo.git` put the token in `operations.log`. Sources are now logged without user info, query or fragment, including where a message repeats them. Thanks to @salmonumbrella for the report. Refs: #878, #881.
|
||||
- **A broken `audit-rules.yaml` no longer turns the audit off** — when the global `audit-rules.yaml` did not parse or one of its patterns did not compile, global scans dropped the error and ran without any pattern rules, so audits silently reported fewer findings. Every scan that uses the rules now reports a scan error that names the file. An install whose audit cannot run is blocked and what it installed is removed, for skills, agents and tracked repositories; `--force` does not override this, and `--skip-audit` still skips scanning. `update` and `upgrade --skill` keep the installed copy, and `update --all` lists the item under "Blocked by security audit" and exits non-zero. Thanks to @salmonumbrella for the report. Refs: #879, #882.
|
||||
- **`install --json` no longer overwrites without `--force`** — `--json` also turned on `--force`, so a script that only wanted JSON output replaced locally edited skills and agents. JSON now changes only the output: existing items are kept and listed in a new `skipped` array, and overwriting needs `--force`. `install -p --json` no longer follows a changed lock pin unless `--force` or `--update` is given. Thanks to @salmonumbrella for the report and the fix. Refs: #877, #884.
|
||||
- **Reinstalling from the same repo is reported as skipped** — the docs say a reinstall from the same repository is skipped with a warning, but the check read only the per-skill file that installs no longer write, so the reinstall failed with "already exists". It now reads `.metadata.json`: reinstalling the same skill of a repo without `--force` exits 0 and lists it under `skipped`. Another skill of that repo that would take the same folder, for example through `--name`, is still a conflict. Reinstalling a single-skill repo with `-s <skill> -a <agent>` still installs the requested agent. In the dashboard, batch installs report kept skills and agents as skipped instead of installed, the install dialog says they are already installed, and a tracked reinstall of the same repo returns 409 instead of an error. The Herdr panel reports skipped names the way it reports failed ones. Refs: #887.
|
||||
- **`skillshare log` shows what an install skipped** — install entries now record skipped skills and agents, shown as `skipped=` in `skillshare log`. This covers CLI installs, `skillshare search` installs, and installs from the dashboard, including config installs. Refs: #886, #887.
|
||||
- **OMP plugins installed into a registered marketplace update again** — installing an Oh My Pi plugin from a marketplace that was already registered used OMP's cached catalog, so OMP recorded an older version and every later update was blocked, while `plugin update --all` skipped it and exited 0. Install now refreshes the marketplace catalog first, and removal follows the version OMP actually recorded once its scope and cache path check out. A target that cannot update is now reported as failed: the other targets still update, `--dry-run` shows the failed rows and exits non-zero, and the plan JSON marks it `failed: true`. Thanks to @salmonumbrella for the report and the fix. Refs: #851, #853.
|
||||
- **Imported plugins no longer fail `plugin update --all`** — the native-support check ran before the rules that intentionally skip a plugin, so an imported Grok plugin made every `plugin update --all` exit non-zero. Intentional skips now stay skips and point to the native client; changes that run a native command are still checked. Refs: #873.
|
||||
- **The Doctor page shows checks as they finish** — the dashboard's Doctor page waited for every check, up to about 30 seconds when a plugin CLI did not answer, and the request was sometimes cut off. Checks now appear as each one finishes, and **Run again** keeps the previous result until the new one is done. The audit, diff and doctor views show the server's reason when a stream fails. API clients can use the new `GET /api/doctor/stream`. Refs: #870.
|
||||
|
||||
## [0.30.6] - 2026-11-11
|
||||
|
||||
### New Features
|
||||
|
||||
Reference in New Issue
Block a user