Reject unsafe subdirs that contain path traversal (..), NUL bytes,
control characters, backslashes, absolute paths, and URL-encoded
traversal (single, double, or overly deep encoding). Validate derived
source names and CloneURLs for NUL and control characters.
SEC003: github.com/owner/repo/../../etc/passwd previously produced a
Source with Subdir containing traversal segments. Now rejected.
TestBackup_AfterSync_ResolvesSymlinks failed ~7-10% of runs. The cause
is in the test, not production: `sync` takes a pre-sync backup of the
target before creating the agent-browser symlink, and the explicit
`backup` takes a second one. When the two land in different clock-seconds
there are two timestamp dirs, and `entries[0]` (lexicographically oldest)
points at the pre-sync backup that predates the symlink, so the
symlinked-skill assertion fails.
Select the newest backup dir (last entry, ascending sort) so the
assertion targets the post-sync backup that actually contains the
resolved symlink. Verified 0/80 failures under -race.
ref #226
Accept leading v prefixes during version checks, keep local metadata builds in dev mode unless they are clean exact release tags, and prevent cleanup warnings from leaking into JSON-mode stderr.
Return errors from branch refresh and checkout when git fetch fails instead of serving stale refs or proceeding with checkout.
Cover branch-list and checkout fetch failures with server regression tests.
Update tracked repo remotes before persisting source metadata so the API cannot report success while the on-disk git remote remains unchanged.
Cover both remote update failure and successful remote plus metadata updates.
Route temp clone cleanup through a shared helper that logs cleanup failures with path and error while preserving existing best-effort cleanup semantics.
Cover cleanup failure logging with a deterministic unit test.
Return explicit comparison errors for non-numeric version segments instead of treating them as zero.
Fail closed in update checks when current or cached latest versions are malformed, and cover numeric, malformed, and cache-preserving cases.
The Dashboard uses POST /api/update (not the stream) and only inspected
res.results.length, so a fresh machine whose only tracked repos are
missing on disk fell through to the "no tracked repos" toast with no
rehydrate hint — leaving one web update-all entry point with the old
issue #212 symptom.
- Type the POST /api/update response with missingTrackedRepos.
- Guard against a null results slice (Go returns nil → JSON null) before
filtering, and only show "no tracked repos" when there is genuinely
nothing — including no missing repos.
- When missing tracked repos are reported, show a warning toast pointing
the user to the Updates page to rehydrate; string added to all 11 locales.
Follow-up to 54dd7d12.
When tracked repo metadata is committed but the gitignored clone dirs are
absent on a fresh machine, `update --all` silently reported zero work: the
batch path skipped missing repos with no message, and the web UI's
update-all paths only listed repos present on disk.
Missing tracked repos are now detected and surfaced everywhere:
- CLI: a warning block (human output) plus an aggregated
`missing_tracked_repos` summary (names + one-shot rehydrate hint) in JSON;
the per-item error is concise. Fixed for both global and project modes,
and batch and single-target paths.
- Server: GET /api/update/missing-tracked-repos and the update-all/stream
summaries report them; POST /api/update/rehydrate re-clones them from
metadata (mirrors bare `skillshare install`).
- UI: the Update page shows a warning banner listing missing repos with a
Rehydrate button; strings added to all 11 locales.
Closes#212
Project-mode agent sync re-reported every agent as updated on the second
sync when the project lived under a symlinked ancestor (e.g. macOS temp
dirs under /var -> /private/var), failing idempotency on macOS CI.
Two interacting causes:
- shouldUseRelative was unstable: on the first sync the target directory
does not exist yet, so evalOrClean fell back to a non-resolved path and
the under-project-root check failed, producing an absolute link; on the
second sync the directory existed and resolved, so it wanted a relative
link and reformatted the existing one. evalOrClean now resolves the
deepest existing ancestor and re-appends the missing tail, so the answer
is the same whether or not the leaf exists.
- The existing-symlink comparison only did a direct string compare, so a
stable relative link whose resolved target carried the unresolved alias
prefix looked like it pointed elsewhere. It now canonicalizes both sides
via EvalSymlinks on the slow path, mirroring isSymlinkToSource.
Adds a regression test that reproduces the symlinked-ancestor case on any
platform.
Factory Droid stores custom droids as markdown with YAML frontmatter in
.factory/droids/, which maps onto skillshare's existing agents sync. Add
the agents path so droids sync alongside skills, and register a factory
alias so the target is discoverable by brand name.
Refs #213
Surface the existing .skillignore glob support in the enable/disable
command reference and the folder-organization guide. Users reported
(issue #209) only finding it buried in reference/filtering, assuming
disable accepted skill names only.
Add "frontend/**" folder examples, a dedicated "Disable a Whole Folder"
section, and shell-quoting guidance.
Add multi-select mode to the Resources page so users can enable or
disable many skills or agents at once (issue #203), instead of toggling
each one individually.
- New POST /api/resources/batch/toggle endpoint (skills + agents) backed
by the shared .skillignore/.agentignore, returning per-item results and
an updated/unchanged/failed summary
- Selection mode across grid, tree, and table views with a unified
checkbox, folder-level tri-state select-all in the tree, a bottom
action bar (Enable applies immediately, Disable confirms first), and
optimistic updates
- i18n keys for all 11 locales
Tier and cross-skill findings are generated programmatically rather than
compiled from the rule list, so they bypassed the disabled-ID filter applied
at rule-compile time. Setting enabled:false in audit-rules.yaml marked the
rule disabled in the listing but the scan still fired it at full severity.
Filter synthetic findings against the disabled-ID set in both the skill-scan
and single-file paths, threading the set through the cross-skill call sites
(analyzer, cmd, server). Fixes#204.
Repeated ** segments caused O(N^K) recursive backtracking where K is
the number of consecutive globstars. Convert to index-based recursion
with (patternIndex, pathIndex) memoization, reducing to O(P*N).
The change is intended to preserve existing matching behavior while
avoiding repeated evaluation of the same (patternIndex, pathIndex)
states. Existing tests pass, and additional globstar semantics tests
were added to guard the core behavior.
Fixes exponential hang on patterns like **/**/**/**/0 with 150+ segment
paths (previously >2s, now ~60ms).