The legacy lookup stat'ed the skill's current entry before looking for old
names, adding syscalls to every flat sync; it now stats only the candidate
names other than the current one (usually none exist) and checks the
current entry only once an old managed entry is found. ReadManifest now
always returns a Naming map, Manifest.Remove drops an entry's records in
one place, copySkillToTarget takes the name to write instead of deciding
it, and diff reads the manifest only for copy targets.
The legacy lookup checks a skill's flat, standard and prefixed names. With
a manifest that has no naming records yet, aaa/dev looked at "dev", found
the top-level dev skill's managed copy and synced into it, because nothing
said the entry belonged to someone else. A skill that already has its
current entry now skips the lookup, and names other skills sync to in the
same run are never treated as old names.
Two tracked repos that ship a skill with the same name cannot both reach a
tool today: standard drops both, flat keeps both folders but tools list
skills by name:. Prefixed gives a skill inside a tracked repo the folder and
name: <repo>-<name> in the copy, so both stay usable and the name shows
where the skill came from. It is copy-only because a merge symlink points
at the source, where name: must not change; config validation rejects it
elsewhere, per target, like extension does.
Switching naming used to know only the flat name as the old one, so
standard <-> prefixed would orphan entries. The migration now checks the
skill's name under every naming, and the copy manifest records the naming
that produced each entry: a switch finds the exact old entry, and a renamed
copy is re-copied so its name: matches even when the source is unchanged.
Manifests without the record fall back to flat, standard, prefixed in turn.
A rename is written to the manifest at once, so a later failure cannot
leave the renamed folder looking user-owned.
The dashboard showed the CLI's raw line, "api-server@claude: include pattern
"*review*" matches no skill in the source", once per target. It named a
config field, did not say the target ends up with no skills, and gave no way
to fix it. Sync now returns these as structured entries; the Sync page
groups a project's targets into one localized notice that says what happens
and links to where the filter is edited. The CLI line also says the filter
adds nothing to the target.
GetTrackedRepos returned OS-native paths, so every caller that compared
them with slash paths (RepoRelPath, dashboard repoPath, metadata keys)
had to convert first. Two callers in install_queries already did, and
this branch added four more. It now returns slash paths on every OS,
matching GetMissingTrackedRepos; filepath.Join and filepath.Base accept
them on Windows, and `update org/_repo` now also matches by full path
there. The per-call-site conversions are gone.
Smaller cleanups in the same code:
- repoOf takes the slash relPath and looks up prefixes by substring, so
the walk no longer splits and joins per directory;
- repo .skillignore matchers are keyed by the same slash rel path, so
isSkillIgnored no longer needs walkRoot;
- batch uninstall uses utils.IsTrackedRepoDir instead of a literal "_";
- the dashboard uses repoOf where it spelled out isInRepo && folderOf;
- the new tests reuse postBatchUninstall and share one --into sandbox.
`install --track --into org` clones to org/_repo, but skill discovery
only marked a skill as in-repo when the first path segment started with
"_". Everything downstream also assumed the repo name was that first
segment. Skills of such a repo were therefore treated as loose skills:
no source in the dashboard, not grouped under the repo, a skill count of
0 in `status`, and an install hint that pointed at the wrong name.
Discovery now records each skill's tracked repo root (RepoRelPath). A
first-segment "_" folder still counts as before; a deeper one counts
only when the walk saw it as a git clone, so a plain org/_drafts folder
stays a folder. The repo's .skillignore is looked up by the same root.
Consumers move from "first segment" to that root:
- the API exposes it as repoPath, and the dashboard groups, updates and
uninstalls repos by it;
- `list` and `status` use it for repo names and counts;
- batch uninstall accepts a nested repo name, so the dashboard can still
remove an --into repo now that its skills are in-repo;
- repo names sent to the dashboard by check and update-all use slash
paths, so they match repoPath on Windows too;
- the install hint suggests `skillshare update org/_repo`.
Fixes#470
canonicalPath relied on filepath.EvalSymlinks, which since Go 1.23 leaves
a Windows junction component in place. A link whose target is itself a
junction, or sits below one, was followed as the unresolved junction:
the walk saw no descendants while the inventory counted as complete, so
target entries from that link could be pruned, and a sync target spelled
through a junction could evade the overlap guard. Every existing path
component is now resolved hop by hop, and a target that still ends in a
link marks the link unavailable instead of being followed.
The policy also kept the source root as spelled, so a relative global
source was discovered but never resolved for writes or audit. Root,
first-level and Resolve comparisons now use the same canonical absolute
form. Copy and checksum roots get the same one-hop resolution so a
followed junction that is itself a skill no longer trips the cycle guard.
Keep the canonical skills source root and logical relative tail when creating or reformatting project links, including Windows relative links and junction fallback. Shared agent and extra links retain full canonicalization, and global Unix link text stays unchanged.
Allow group walks outside the source only for an enabled first-level link validated by Follow.Resolve, remapping children to logical paths and metadata keys while retaining nested-link guards and root-skipping semantics. Resolve copy and checksum operation roots before walking to avoid false cycles, and preserve followed source links during forced collect with per-item warnings.
Regression coverage verifies second-sync ownership, status, prune, unlink, format changes, linked-root copying, group guards, and forced collect preservation. Linux build, vet, sync and command tests, sourcewalk ratchet, server collect tests, context validation, and Windows cross-build passed; Windows runtime was not verified. No new raw-walk allowance was needed.
Users link a checkout directly under the skills source (e.g.
skills/_dev-skills -> ~/code/dev-skills) and expect it to behave like any
other first-level entry. sourcewalk resolved only the root, so these links
were invisible to discovery, sync, update, and the dashboard.
follow_source_links (global and project config, default false) enables a
per-operation sourcewalk.Follow policy. Walk, WalkDir, and ReadDir report a
followed link under its own name as a directory and descend into its
target, so names, RelPath, FlatName, SourcePath, tracked-repo detection, and
metadata keys stay logical. Only entries directly under the source root are
followed. The policy is built from config and passed explicitly to every
discovery caller in the CLI, server, and reconcile, so counts agree; the
server keeps a per-instance snapshot refreshed on config load and save
instead of a process-global flag.
A link is skipped, with one warning per operation, when its target is
missing or unreadable, is the source or a parent of it, or overlaps an
active skills target. A missing or unreadable target marks the operation
incomplete: sync then skips pruning and copy-orphan deletion, and reconcile
keeps metadata, so an unmounted drive or a re-clone in progress cannot
erase target entries or install records. A link to an existing file is not
a directory link and stays an ordinary entry; treating it as incomplete
would disable pruning for anyone sharing a file such as .skillignore.
doctor keeps raw link identity: with the key off it still says the link is
not followed and names follow_source_links: true; with it on it reports
followed links and skipped links with the reason.
The raw-walk allowlist gains entries for the policy path in Walk and
WalkDir: the source root walk with a policy, and the walk of a validated
link target whose paths are remapped to the logical link path. Both live
in sourcewalk itself, the package that owns source traversal.
Refs #274
The unmatched include warning from #326 was written straight to
DiagOutput. The dashboard sync never showed it because DiagOutput is the
server's stdout, and an interactive CLI sync could redraw its progress
area over it. Return the warnings with the merge and copy results so
SyncSkillTarget adds them to the target's warnings, which both the CLI
and the API already report. Prefix them with the target name, since the
API returns one list for all targets.
A top-level skill name has no path separator, so the suggestion cut off
its first character and suggested "alpha" for a typo like "lpha". Compare
the whole name in that case.
Refs #325
Target include/exclude patterns are matched with filepath.Match against the
source-path flat name, while target_naming: standard shows the bare SKILL.md
name in the target directory. A pattern written with the name the target shows
selects nothing, so sync prunes the entries a working pattern had linked, and
the only trace is the echoed include: line, which never says that any of those
patterns matched zero skills.
Add FindUnmatchedIncludes and render one diagnostic per include pattern that
selects no discovered skill, suggesting the flat names whose last path segment
is the pattern. exclude is left alone: an exclusion that blocks nothing is the
normal state of a spare filter, while an include that selects nothing is always
a mistake and is the destructive direction. The diagnostic goes to DiagOutput,
so --json output is unchanged, and no filtering, naming, or pruning behavior
moves.
Closes#325
Since #315, a live link into the source is pruned only when the manifest
tracks it. The ManagedOnly comments still said any link into the source
counted as skillshare's, which no longer matches what the leftover-dir
cleanup removes.
Merge-mode prune removed any target entry that was a symlink or junction
resolving under the skills source, without checking whether skillshare had
created it. Links made by hand or by another tool were therefore deleted by
name, silently and outside trash, so they were not recoverable through
restore.
Ask the manifest, which the directory branch of this same function and all
of copy-mode pruning already treat as the ownership record: a live link into
the source is pruned when it is tracked there or when --force is set, and is
kept and reported as a local entry otherwise. Broken links into the source
and the force escape hatch behave exactly as before.
Some tools only need skillshare for AGENTS.md, MCP or agents, while their
skills folder is managed elsewhere or shared through another target.
Syncing skills into those folders duplicated what the tool already reads,
so a target can now turn skills off (`target <name> --no-skills`, the
dashboard switch) and sync, diff, status and doctor skip its skills.
Switching off warns about the other tools that read the same folder.
The target list also gains tools that were missing, with colored vendor
logos where their licenses allow it, and zencoder/replit default paths
follow their current documentation.
The dashboard's target pages get the matching UI: universal is pinned as
the shared folder when adding a target, the add dialog lists what will be
written, and the AGENTS.md tab shows where the file comes from in one card
with an in-place switch between shared files.
Since Go 1.23, os.Lstat reports NTFS junctions as irregular files instead
of symlinks (winsymlink, active for go 1.25 in go.mod). Every check that
looked only at os.ModeSymlink therefore missed the junctions skillshare
itself creates on Windows: linked skills could be treated as local folders,
backed up, copied through, or left behind on prune.
Detect links through utils.IsLinkMode, which also asks the OS about the
reparse tag of irregular entries, and count only real symlinks and
junctions as links so OneDrive placeholders and deduplicated files stay
regular files. restore.go also used os.Stat for its link check, which
follows the link and never matched.
Refs #299
Flat target naming turns a source skill under a dot directory
(e.g. .system/example) into a hidden target entry (.system__example).
Target scans skipped every hidden name, so diff, doctor, status and the
dashboard reported these synced skills as missing, and prune never
removed them after the source skill was deleted.
Hidden entries are now skipped only when the target manifest does not
record them as managed, so files like .skillshare-manifest.json or
.DS_Store are still ignored.
Refs #294
The analyze numbers were drifting from what tools really see. Symlink-mode
targets expose the whole source folder, so .skillignore'd skills still load
but were left out; merge/copy targets often hold skills the user dropped in
by hand that skillshare never counted; and chars/4 undercounted CJK text by
roughly four times.
Route the CLI, TUI, web handler and sync summary through one
sync.TargetSkills so the three copies of the filter logic cannot disagree
again. Disabled skills stay in the discovery with Disabled=true and are
priced for symlink targets; unmanaged folders in a target are reported as
Local. Token estimates now charge one token per wide rune.
The dashboard keeps the chosen target in the URL so target and project
pages can deep-link into the analysis, lists project targets after own
targets with their tool icon, and tags local/disabled rows. Select gives
single-select icons a key to silence a React warning.
Both tools now document ~/.agents/skills and <project>/.agents/skills as
their skills locations and keep the old directories only for backward
compatibility. Writing to the old path next to `universal` makes them read
every skill from two places.
Global configs pin absolute paths, so they are unaffected and keep their
overlap warning. Project configs store target names only, so an existing
project follows the new default on its next sync while the links it created
in .goose/skills or .openhands/skills would stay behind and show up as
duplicates. Project sync now removes those leftovers from also_scans
directories that no configured target writes to.
That directory belongs to another tool, so the cleanup runs the prune with
ManagedOnly: only symlinks into the source and manifest-tracked folders are
removed. The flat-name heuristic would otherwise delete hand-made folders
such as `_drafts`. Targets with an explicit path are skipped, and --dry-run
only previews.
Refs #135
Apply config ignore patterns to copy-mode file traversal, checksum, mtime, CLI diff, and server sync/diff paths.
Add default ignores for .DS_Store, .git, and __pycache__ in global and project init configs.
Tests: go test ./... -count=1
DiscoverSourceSkillsWithStats did not set collectContext=true, so
DescChars/BodyChars were always 0 when sync called buildAnalyzeEntries.
Add DiscoverSourceSkillsWithStatsAndContext that enables both ignore
stats and context collection in a single walk, and use it in sync.go
and sync_project.go.
Two correctness improvements from Codex review:
1. EvalSymlinks before relative-link decisions: shouldUseRelative and
createLink now resolve symlinked ancestors via EvalSymlinks so that
a project opened through a symlinked workspace, or a target dir that
traverses a symlink, produces correct relative paths (or falls back
to absolute when paths diverge from the project root).
2. Atomic reformat via reformatLink: instead of remove→create (which
loses the link on createLink failure), reformat now creates a temp
link and renames it over the original. On Unix this is atomic; on
Windows it falls back to remove→create when rename fails.
Tests added for symlinked project root, divergent target symlink,
and reformatLink atomic swap.
filepath.Abs(dest) on a relative readlink result resolves against CWD,
not the symlink's parent directory. This broke the 'already correct'
check when comparing relative symlinks to absolute source paths.
Add resolveReadlink() helper that joins relative targets with
filepath.Dir(linkPath) before calling filepath.Abs.
Also check os.Remove errors in all reformat paths instead of ignoring
them silently.
Addresses PR #122 review feedback.
sync previously skipped symlinks that pointed to the correct source,
even if their format (absolute vs relative) didn't match the expected
mode. This meant existing absolute symlinks in project mode were never
converted to relative on subsequent syncs.
Add linkNeedsReformat(dest, wantRelative) helper that checks whether
the raw readlink result has the wrong format. All four 'already
correct' skip points now check format and recreate when mismatched:
- SyncTargetMergeWithSkills (skills merge mode)
- SyncTarget (skills symlink mode)
- syncExtraSymlinkMode (extras directory symlink)
- syncOneExtraFile (extras per-file merge)
Global mode is unaffected: shouldUseRelative returns false, so
linkNeedsReformat never triggers on absolute links.
Also fix macOS /var → /private/var path mismatch in
TestSyncProject_RelativeSymlinks by EvalSymlinks on expected path.
SyncTargetMerge, SyncTargetMergeWithSkills, CreateSymlink, and SyncTarget
now accept projectRoot. When non-empty, createLink uses relative symlinks
for paths under the project root.
All existing callers pass empty string (global mode, no behavior change).
Ref: #120
Unix: computes filepath.Rel from link dir to source, falls back to
absolute on error.
Windows: tries os.Symlink with relative path first (Developer Mode),
falls back to mklink /J with absolute paths.
All existing call sites pass false (no behavior change yet).
Ref: #120
- Enable frontmatter parsing in DiscoverSourceSkillsAll so /api/skills
returns saved targets (fixes target display reverting to All)
- Skip tracked-repo skills in batch/single target handlers to prevent
dirty git state; reject with 400 for single-skill endpoint
- Add RefreshMetaHashes to recompute audit hashes after SKILL.md edits;
batch handler refreshes outside the write lock
- Show 'Uninstall Repo' with confirm dialog for tracked-repo members
instead of the always-failing individual uninstall action
- Return 409 Conflict when enabling a skill disabled by glob pattern
in .skillignore/.skillignore.local instead of false success
- Preserve IsInRepo flag for root-skillignore-disabled skills so they
remain visible in the Tracked tab after disable
- Update changelog and website changelog for v0.18.7
With 25000+ skipped skills (standard naming + name collisions),
the CLI sync output was flooded with per-skill warning lines.
Now prints a concise summary:
12345 skill(s) skipped (naming validation)
456 name collision(s) excluded
Applied to both merge and copy sync modes.
Add per-target target_naming setting that controls how skill directories
are named in sync targets. 'flat' (default) preserves the existing
flattened __ separator behavior. 'standard' uses the SKILL.md name field
directly, following the Agent Skills specification.
Standard mode:
- Validates SKILL.md names match directory basenames
- Detects and skips name collisions with warnings
- Auto-migrates managed entries when switching from flat to standard
- Preserves legacy entries when bare name is already occupied
Also includes:
- Per-target toggle via CLI (--target-naming), TUI (N key), and Web UI
- Collision output redesigned as deduplicated compact summary
- Config inheritance: global target_naming → per-target override
- JSON schema + UI validation for the new field
Closes#110
New discovery function that includes .skillignore-d skills marked with
Disabled=true. Existing discovery functions are unchanged — Disabled
defaults to false for backward compatibility.
Populate LintIssues during skill discovery (collectContext path) and
propagate them through analyzeSkillEntry so JSON output includes lint
findings per skill.
Press R on a selected target to show a confirmation overlay.
Confirming exits the TUI and dispatches the full target remove
flow (backup + unlink + config update), matching the existing
exit-with-action pattern used by the skill list TUI.
Supports both global and project modes.
- Remove early return in dry-run so linked/copied counts are populated
correctly (P1: UI preview no longer hides Sync Now button)
- Use local quietDryRun flag instead of mutating global DiagOutput
to suppress per-skill messages (P2: safe under parallel sync)
- Use "Will create" for dry-run, "Created" for actual sync (P3)
- Include dir_created in allUpToDate check so empty source + missing
dir still shows Sync Now in the preview modal
v0.17.5 blocked sync when target directories didn't exist, which
prevented first-time users on fresh CLI installs from syncing.
Sync now auto-creates missing directories and reports what it did
in CLI output, dry-run preview, and the Web UI. Init also creates
target directories when it detects an installed CLI tool.
Web UI: Sync Preview modal stays open after sync with a success
banner instead of immediately closing, so users can review results.
Closes#87
Unify ~160 lines of duplicated walk logic from DiscoverSourceSkills and
DiscoverSourceSkillsLite into a single discoverSourceSkillsInternal
function controlled by discoverOptions (parseFrontmatter, collectIgnored,
collectTracked).
Add DiscoverSourceSkillsWithStats which returns IgnoreStats populated
with root/repo .skillignore file paths, active patterns, and ignored
skill relPaths. When collectIgnored is true, CanSkipDir is bypassed so
the file-level Match check can record every ignored SKILL.md path.
Add Patterns() method to skillignore.Matcher to expose the original
pattern strings stored during Compile().
Replace the naive string matcher (~40 lines) with a proper gitignore
implementation (~200 lines, no external deps) supporting **, ?, [abc],
!negation, /anchored, pattern/ (dir-only), and \#/\! escapes.
New Matcher API (Compile, ReadMatcher, Match, CanSkipDir, HasRules)
replaces the old Match/ReadPatterns functions (kept as deprecated
wrappers). Call sites in sync.go and install_discovery.go updated.
Also fixes isDir=true for install discovery (skill paths are dirs).
100% test coverage on skillignore package (~150 table-driven cases).
Added 7 sync discovery tests and 5 install integration tests covering
**, !negation, pattern/, [Tt] char class, and combined patterns.
Skip entire ignored directory trees during walk instead of filtering
at SKILL.md level. Support root-level .skillignore at source root
for globally muting skills or excluding patterns.
Root patterns checked first (can skip entire tracked repos), then
repo-level patterns. Both layers use filepath.SkipDir for performance.
Ref: #83
- Extract duplicate .skillignore filtering block into isSkillIgnored()
- Use utils.IsTrackedRepoDir consistently in DiscoverSourceSkillsLite
- Fix unchecked os.WriteFile in skillignore test
DiscoverSourceSkills and DiscoverSourceSkillsLite now load .skillignore
from tracked repos and skip matching paths during walk. This prevents
false warnings in doctor/status/list for vendored SKILL.md files
(e.g. .venv/ inside a tracked repo).
Extract shared skillignore package from internal/install so both
install-time and source discovery use the same filtering logic.
copyDirectoryWithState did not skip .git/ dirs, causing collect to fail
on git-cloned repos (e.g. obra/superpowers) — filepath.Walk would abort
on .git pack files, leaving only empty directories in source.
Add copyDirectorySkipGit() variant and use it in PullSkill. Existing
copyDirectory (used by copy-mode sync) is unchanged.
Keep sync dry-run diagnostics on stdout by default to avoid non-JSON behavior regressions. Introduce sync.DiagOutput so JSON mode can silence diagnostics via io.Discard at command entry and restore after execution. Update search JSON integration assertion to match fully silent progress behavior.
Address code review feedback from PR #67:
- Move all dry-run diagnostic messages (internal/sync) from stdout to
stderr so they don't pollute JSON output
- Suppress UI output (spinners, headers, warnings) in uninstall --json
mode by adding jsonOutput guards across all phases
- Return nil after writeJSONError to prevent dual output (JSON + plain
text error) in install and sync commands
- Handle json.MarshalIndent error in writeJSONError with stderr fallback
- Merge duplicated discovery logic (JSON vs non-JSON branches) in sync,
sync_project, diff, and collect using nullable spinner pattern
- Remove checkRepoDirty alias in status.go, call isRepoDirty directly
- Remove extractJSON test workaround — tests now assert stdout is valid
JSON directly, enforcing the clean-stdout contract
filepath.Walk/WalkDir silently skip symlinked root directories because
os.Lstat sees the root as a non-directory symlink and never enters it.
This affected update, uninstall, reconcile, and server handlers when
the source directory was a symlink (common with dotfiles managers).
Extract ResolveSymlink() from sync.go into internal/utils/path.go and
apply it at every filepath.Walk/WalkDir callsite on source/target dirs.
Also add containment guards to resolveGroupSkills (uninstall) and
resolveGroupUpdatable (update) to reject group directories that are
symlinks pointing outside the source tree.
When source or target directories are symlinks managed by external tools
(e.g., dotfiles managers), sync would incorrectly delete them during
mode conversion. This change:
- Add resolveWalkRoot() so filepath.Walk follows symlinked source dirs
- Add isSymlinkToSource() with canonicalization to distinguish internal
symlink-mode targets from external symlinks (alias-safe)
- Protect external target symlinks in both merge and copy mode
- Allow FindLocalSkills to scan through external symlinked targets
- Handle external symlinks in CheckStatusMerge as normal directories
When a target directory is an external symlink (e.g., dotfiles manager
pointing ~/.claude/skills/ elsewhere), sync previously deleted it
unconditionally during symlink→merge/copy mode conversion. This also
affected DiscoverSourceSkills which failed to traverse symlinked source
directories because filepath.Walk uses Lstat on root.
Changes:
- Add resolveWalkRoot() helper; use in both Discover functions so
filepath.Walk enters symlinked source directories
- Add isSymlinkToSource() to distinguish source-pointing symlinks
(symlink mode residue, safe to remove) from external symlinks
- Pass explicit sourcePath to SyncTargetMergeWithSkills and
SyncTargetCopyWithSkills for reliable symlink origin detection
- Fix FindLocalSkills to use utils.ResolveLinkTarget (handles relative
symlinks correctly) and follow external symlinks for scanning
- Fix CheckStatusMerge to follow external symlinks instead of returning
StatusConflict
- Add 17 unit tests and 6 integration tests covering symlinked source,
target, both, chained symlinks, idempotency, prune, and pull