Improve all 5 active skills with pushy descriptions for better
triggering accuracy, following devcontainer as the gold standard.
- changelog: add trigger scenarios (release notes, version prep)
- cli-e2e-test: add trigger scenarios (runbook, E2E validation)
- codebase-audit: add 3 new audit dimensions (handler split, oplog
coverage, web API consistency) and trigger scenarios
- implement-feature: add project patterns reference (handler split,
dual-mode, TUI, web API, oplog) and trigger scenarios
- update-docs: expand scope to all 7 doc categories with code-to-doc
mapping guide and trigger scenarios
Also remove empty .skillshare/skills/frontend/ directory.
Duplicate findings (same pattern+file+line+snippet) are noise.
Global deduplication is now the default for all profiles.
Use --dedupe legacy to opt back into the old behavior.
config.Load() normalized empty block_threshold to "CRITICAL", preventing
ResolvePolicy from applying profile-based thresholds (e.g. strict → HIGH).
Change: normalizeAuditBlockThreshold("") now returns "" instead of "CRITICAL".
All downstream callers (install, update, server) already use
audit.NormalizeThreshold() which handles empty → CRITICAL independently.
Also adds 4 integration tests for --profile, --dedupe, and config-based
profile resolution. Includes gofmt formatting fixes.
P1: mergeYAMLRules Phase 2 used full-object replacement for overlays
without Enabled set, causing severity-only entries ({id, severity}) to
wipe regex/pattern/message from the base rule. compileRules then failed
with "empty regex". Now uses field-level merge: only entries with Regex
do full replacement; otherwise Enabled and Severity are patched
individually.
P2: TogglePattern enable only removed pattern-level disable entries,
leaving ID-level {id: "x", enabled: false} untouched — both builtin
and custom rules. Now clearDisableForPattern matches by builtin ID set
OR by r.Pattern == pattern, covering custom rules in the same group.
pterm's cursor package writes hide/show ANSI codes (\x1b[?25l/h)
directly to os.Stdout, bypassing WithWriter(). The previous fix
(SetProgressWriter to stderr) was insufficient because pterm still
created progress bar objects that triggered cursor.Hide()/Show().
Add SuppressProgressToStderr() which both redirects ProgressWriter
and forces isProgressTTY() to return false, preventing pterm
components from being instantiated at all. Extract assertNoANSI
test helper and cover all three structured formats (JSON, SARIF,
Markdown).
Also update pre-commit E2E runbook with steps 10-11 (framework
integration tests) and add pre-commit-hook recipe to sidebar.
- Add --format markdown for self-contained audit reports (GitHub Issues/PRs)
- Simplify pterm box rendering: replace DefaultBox with separator-line style
- Remove unused SkillBox, SummaryBox, SkillDisplay, wrapText helpers
- Color Box titles cyan (consistent with HeaderBox)
- Add breathing room: blank lines around progress bar and after summary
- Show "skillshare audit (project)" header title in project mode
- Append elapsed time to cross-skill synthetic results to fix slice mismatch
- Add --format <text|json|sarif> flag to audit command
- Deprecate --json in favor of --format json (with stderr warning)
- SARIF output uses ToSARIF() from internal/audit/sarif.go
- Suppress UI (spinners, progress bars) for all structured outputs
- Update auditPath/auditSkillByName to accept format string
- Update help text with new flag and examples
- Add 3 integration tests: FormatSARIF, FormatJSON, JSONDeprecation
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 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
- Extract shared applyDetailScroll into tui_helpers.go (was duplicated
across diff_tui.go, log_tui.go, audit_tui.go)
- Create utils.FileHash and utils.FileHashFormatted, replacing private
copies in install/meta.go and audit/audit.go
- Export install.MetaFileName constant, use from diff_file.go
- Replace double SHA-256 in fileContentsEqual with streaming byte compare
for early-exit efficiency
- Add 1MB size cap to generateUnifiedDiff to prevent performance issues
- Move noTUI into diffRenderOpts struct, simplifying function signatures
- Remove dead field diffTUIModel.results (never read)
- Replace hand-rolled countOccurrences with strings.Count
- Skip latestMtime() walk for merge mode (symlinks make it misleading)
- Fix ensureFiles() nil cache miss (return non-nil empty slice)
- Cache sorted items + categories in TUI model (avoid per-render recompute)
Interactive two-phase TUI: target picker → version list with left-right
split panel layout. Includes detail panel with skill diffs, descriptions,
and target state. Wires no-args restore to TUI dispatch with --no-tui
fallback.
Add ListBackupVersions() and ListTargetsWithBackups() for TUI timestamp
selection. Fix merge-mode targets producing empty backups by following
top-level symlinks in copyDirFollowTopSymlinks().
Add bubbletea-based TUI with multi-select, fuzzy filter, and inline
restore/delete/empty operations. Follows the same dispatch pattern as
list and log: JSON → TUI (TTY + items + !--no-tui) → plain text.
Follow-up to PR #58: check_project.go now extracts custom target names
from project config and passes them to findUnknownSkillTargets, matching
the global-mode fix. Also surfaces config load errors as warnings instead
of silently ignoring them. Adds integration tests for check and doctor
(global + project mode) verifying no false "unknown target" warnings.
The skip-unchanged optimisation (ec81fc1) causes update --all to skip
skills whose metadata already matches the repo state, producing no
InstallResult and therefore no audit output. The test now invalidates
one skill's metadata between install and update, forcing a re-install
that exercises the audit code path.
- Change universal description from "(recommended)" to "(shared agent directory)"
in detect, checklist, and discover flows
- Add FAQ section on using universal alongside npx skills: merge vs copy mode
differences, name collision caveats, lock file visibility
- Add universal info admonition to supported-targets page
- Add E2E runbook for universal target path verification
- Update CHANGELOG and release notes for v0.16.4
The universal target pointed to ~/.config/agents/skills (wrong) instead
of ~/.agents/skills, causing `npx skills list` to miss synced skills.
Also auto-recommend universal as a target whenever any CLI is detected
during `init` and `init --discover`, since ~/.agents/ won't exist on
disk until skillshare creates it.
Closes#54
After pre-flight filters out dirty tracked repos, recalculate summary
so spinner and confirm prompt use correct noun type. Print skip count
when repos are skipped. Replace all "(s)" patterns with pluralS() for
natural English output ("2 groups" not "2 group(s)"). Add isMixed()
method to avoid magic string comparison on noun().
Switch config install and search install UI from plain spinner to
TreeSpinner with StepResult summary showing installed/skipped/failed
counts and elapsed time. Add git clone progress callback to spinners
so users see real-time cloning status.
- Add Quiet field to InstallOptions for suppressing per-skill output
- Add Skipped field to ConfigInstallResult for accurate summary
- Update integration test assertion to match new summary format
Use a repo-root source in grouped search installs by normalizing from CloneURL instead of carrying subdir Raw values. This prevents malformed metadata sources (for example duplicated subdir paths) and keeps provider-agnostic behavior for GitHub/GitLab/Bitbucket/Azure and SSH URLs.
Also centralize discovery source formatting for metadata so whole-repo SSH installs use //subdir syntax consistently.
Add regression coverage:
- cmd unit tests for grouped clone source normalization across providers
- install unit tests for discovery source formatting (including SSH)
- integration test in tests/integration that verifies grouped install metadata source round-trips through ParseSource.
runLogStats() only checked the --audit flag to pick the log file,
ignoring filter.Cmd. This caused `log --stats --cmd audit` to read
operations.log and report "No log entries" even when audit data existed.
Align with runLog() by also checking f.Cmd == "audit".
Replace cryptic internal terminology (+/~/- source only, local copy, etc.)
with action-oriented messages that tell users what to do next (→ sync will
add, ⚠ local copies, ← collect will import). Add overall summary line at
top ("N targets need sync, N fully synced"). Show conditional hints that
only mention --force when local copies actually exist. Use full command
names (skillshare sync) in hints.
When multiple targets produce the same diff output (e.g. all missing
the same skills), they are now merged into a single block with a
combined header like "claude, agents" instead of repeating the full
list for each target. Fully synced targets are also consolidated
into one line (e.g. "codex, copilot: fully synced").
Refactored show*Diff → collect*Diff to separate data collection
from rendering, enabling cross-target fingerprint grouping via
renderGroupedDiffs().
Copy mode previously computed DirChecksum (Walk + ReadFile + SHA256)
for every skill on every sync, even when nothing changed. Add mtime-
based short-circuit: if source dir max mtime matches the manifest
record, skip checksum entirely.
Changes:
- Add DirMaxMtime() — Walk + lstat only, no file reads
- Extend Manifest with Mtimes map (omitempty, backward-compatible)
- SyncTargetCopyWithSkills: mtime fast-path with target validity check
- diff: two-phase scan-then-display with spinner (no more interleaved
spinner + diff output), discovery spinner for both global/project
- Clearer diff reasons: "missing" → "source only",
"missing (deleted from target)" → "deleted from target"
- Server handleDiff: same mtime fast-path for copy mode
- Prune cleans up orphan mtime entries
DiscoverSourceSkills was called 1+2N times (N = number of targets),
each doing a full filepath.Walk + SKILL.md frontmatter parsing.
Now discovered once and passed to all targets via WithSkills variants.
Changes:
- Add SyncTargetMergeWithSkills, PruneOrphanLinksWithSkills (PruneOptions struct)
- Add SyncTargetCopyWithSkills with onProgress callback, PruneOrphanCopiesWithSkills
- Discovery spinner + per-target reporting + summary line in CLI
- Copy mode shows per-skill spinner progress during slow checksum+copy
- Collision report uses dim gray text to avoid confusion with sync results
- Server handlers (handleSync, handlePull) use single discovery pass
- Fix pre-existing TestLog_StatsOutput assertion ("Success rate" → "OK:")
- Report "partial" status for sync/install/update/uninstall when some
items succeed and others fail
- Instrument 6 new commands: check, diff, trash, init, upgrade, and
all mutating commands now log to operations.log
- Auto-truncation via WriteWithLimit with configurable log.max_entries
(default 1000) and 20% hysteresis to avoid truncating every write
- Stats engine: computeLogStats with per-command breakdown
- CLI --stats flag with colored output and ✓ok/total ratios
- TUI stats panel (toggle with 's'): bar charts, status distribution,
last operation timestamp
- TUI stats footer: compact always-visible summary line
- TUI detail panel: j/k scrolling for long entries, proportional
list/detail layout (40%/60%)
- Web API: GET /api/log/stats endpoint
- Updated website docs for log command
- update --prune: detect skills deleted upstream and move them to trash
(batch + single-target paths), clean registry entries after prune
- update without --prune: warn about stale skills with suggestion
- check: report "stale" status for skills whose subdir no longer exists
in the remote repo (previously misreported as update_available)
- Refactor updateTrackedRepo, updateRegularSkill, updateTrackedRepoQuick,
updateSkillFromMeta to accept *updateContext instead of 7-10 individual
params; delete dead updateSkillOrRepo function
- Add integration tests for prune (standalone, nested, all-stale,
registry cleanup) and project-mode prune
- Update website docs for update --prune and check stale status
Add missing replit badge + id anchor in supported-targets.md.
Add project-mode integration tests for audit, check, diff, and log
commands to close gaps found in codebase audit.
Skills are now stored in registry.yaml, not config.yaml. Update 10
integration tests to read from the correct file:
- install_group_test.go: 3 tests read registry.yaml after install
- uninstall_test.go: 3 tests read registry.yaml after uninstall
- uninstall_project_test.go: 4 tests write skills to registry.yaml
(project uninstall doesn't trigger config migration) and verify
registry contents after uninstall
Cover: latest backup round-trip, specific timestamp selection,
target not in backup, dry-run with force, force overwrite on
non-empty target, and no-force failure on non-empty target.
Previously `skillshare check` compared repo-level commit hashes, causing
all skills from a registry repo to show "update_available" whenever any
commit was pushed — even if the individual skill's files were unchanged.
Now each skill records its git tree hash (subdir-level SHA) at install
time. During check, a three-tier comparison is used:
1. Fast path: commit hash matches → up_to_date (zero extra cost)
2. Tree hash path: blobless fetch (~150KB) + batch ls-tree → per-skill
comparison; unchanged subdirs stay up_to_date
3. Fallback: no tree hash (old installs) → commit comparison (unchanged)
Also fixes pterm progress bar printing twice when it auto-stops at 100%.
groupSkillsByDirectory() used filepath.Dir() which always returned the
same parent regardless of recursion depth. Add a prefix parameter so
each level strips already-navigated directories before grouping.
Also fix stale assertion in TestCheck_RegularSkill_ShowsMeta to match
the summarized local-skill output introduced in 37f3425.
Return sizeBytes (int64) instead of sizeMB (float64) from backup API
so the frontend controls display formatting. CLI backup commands now
use the existing formatBytes() helper for human-friendly output.
Extract duplicate formatSize() from CollectPage and TrashPage into
shared ui/src/lib/format.ts utility.
Replace basic text output with bubbletea-based interactive TUI for the
list command. Falls back to plain text with --no-tui or non-TTY output.
- Bubbletea TUI: fuzzy filtering, keyboard navigation, detail panel
- Detail panel: description (SKILL.md), disk path, files, synced targets
- Search: `skillshare list <pattern>` filters by name/path/source
- Filter: `--type git|local` to filter by installation type
- Sort: `--sort name|date|type` with mode-specific defaults
- JSON: `--json` for CI/scripting consumption
- Pager: less/more wrapper for --no-tui plain text output
- 29 integration tests covering all new functionality
Swap fast path order: sparse checkout is now fast path 1 (works for all
hosts including file://), GitHub Contents API is fast path 2 (fallback
when sparse is unavailable or fails). Also reset subdirPath on sparse
failure to ensure proper fallback chain.
Fix red team helpers to inject -g flag after subcommand, preventing
.skillshare/ auto-detection from hijacking global-mode tests.
Add blocked section assertions to TestUpdate_BatchAll_FailsOnMalicious.
Add audit_output_antigravity_runbook.md covering install --all and
update --all audit output verification against sickn33/antigravity-
awesome-skills (7 steps, all passing).
Fix audit_output_online_test.go: batch install exits 0 when some
skills succeed (blocked count is a warning, not an error), so
AssertFailure should be AssertSuccess.
UpdateSkillsFromRepo was using filepath.Join(sourceDir, meta.Subdir) to
derive the local destination, but meta.Subdir is the repo-internal path
(e.g. "skills/foo") which may differ from the local install path ("foo/").
This created leaked copies at sourceDir/skills/foo/ alongside the correct
sourceDir/foo/, doubling the skill count on subsequent scans.
Change skillTargets from []string to map[string]string where key is the
repo subdir (for discovery lookup) and value is the local absolute path
(provided by the caller who already knows where the skill is installed).