Overriding the audit gate with --force was per-invocation, so skills that
legitimately quote attack strings (security scanners, red-team docs) had to
be forced on every `update --all`. The only alternative, disabling the rule,
was global and turned it off for every skill.
Record the accepted findings (rule + file + matched text, no line number) in
.metadata.json under `audit_accepted`, keyed by skill relPath so it survives
entry rewrites and works for tracked repos. Later updates mark matching
findings as acknowledged and exclude them from the block decision; any new
finding, or the same rule matching different text, still blocks.
Staged installs audit a temp directory, so InstallOptions gains an
AuditAcceptRoot/AuditAcceptPath override pointing at the final destination.
Refs #279
Avoid discarding all file hashes when a directory symlink cannot be read as a file. Preserve file symlink hashing and error propagation for broken links and other hash failures.
Add regression coverage for directory symlinks, file symlinks, and broken links. Verified with make check in the devcontainer.
Refs #272
Source parsing only treated inputs starting with /, ~, ./ or ../ as local
paths, so Windows paths such as D:\folder fell through to git URL parsing
and failed with "unrecognized source format" (or were expanded into a
bogus GitHub shorthand when written with forward slashes).
Refs #269
Skills installed from a non-default branch record that branch in metadata,
but the grouped update path (update --all) cloned the remote default branch,
so the skills were either downgraded or reported as removed upstream and
pruned. The web UI update and check handlers had the same gap, comparing
against the remote HEAD instead of the installed branch.
Group batch updates and web UI checks by URL+branch, pass the branch through
to the clone, and resolve remote tree hashes for that branch as well.
Refs #268
The grouped update path (update --all) removed the existing skill directory
before copying the new content, and the audit-block cleanup then removed the
freshly copied content as well. A rejected update therefore deleted the
previously installed, already-audited version and left a stale metadata
entry that update --all could no longer discover.
Stage the new content in a temp directory when updating, audit it there, and
only swap it into place once the audit passes, matching the single-skill
update path.
Refs #271
InstallOptions.Force carried two meanings: "overwrite what is already
there" and "proceed despite audit findings". Update paths set Force
unconditionally to overwrite a staged temp dir, so they had to leave it
off to keep the audit gate alive -- which swallowed the caller's --force
and made the gate impossible to override from either the CLI or the web
UI.
Split the audit intent into InstallOptions.AuditOverride and route every
gate through it alone. Entry points translate --force explicitly, so a
missed one degrades to "force has no effect" rather than "audit is
bypassed". Scan failures stay fail-closed regardless: accepting known
findings is not the same as being unable to learn any.
This also closes two paths that bypassed the gate outright, since both
inherited Force for overwrite reasons alone: install --json and grouped
batch updates.
In the web UI, Force Retry now appears only on failures force can
actually resolve -- a permission-denied unlink previously offered a
button that could never succeed -- and CLI-only flag hints are stripped
from messages rendered there.
Behaviour change: a --force at install time no longer exempts later
updates from the audit gate; it is a one-off decision per command.
Discovery skips well-known target dot-dirs (.claude, .cursor, ...) so that
target-synced copies are not mistaken for source skills. filepath.Walk visits
the walk root first, so pointing an install directly at such a directory —
`skillshare install user/repo/.claude` — matched the guard on the root itself
and returned SkipDir before scanning anything, yielding zero skills.
Exempt the walk root from the dot-dir filter in discoverSkills and
scanAgentDir: it was requested explicitly, so filtering it can only discover
nothing. .git stays unconditionally skipped, and repo-root walks are
unchanged, preserving the original guard.
Batch update looks up each skill's subdir in the discovery map, but
discovery deliberately skips target dot-dirs (.claude, .codex, ...) to
avoid counting target-synced copies as source skills. A skill installed
from such a subdir therefore never appears in the map and is falsely
reported as stale (deleted upstream), steering users toward --prune on a
healthy skill. Fall back to a direct SKILL.md stat in the cloned repo --
mirroring the single-skill update path -- before declaring the path
missing.
Fixes#261
Retry generic HTTPS clone failures by progressively treating additional path segments as part of a nested GitLab repository. Limit retry to repo-boundary errors and preserve the final fallback error for debugging.
The project-root self-install guard lived only in the CLI handler, so the
web UI could still copy the project root into its own .skillshare/skills/
subtree in project mode. Extract the check into a shared
install.RejectProjectRootLocalInstall and apply it in handleInstall and
handleInstallBatch, returning 400 with the same guidance.
Also restructure the rejection message into a cause line plus an
actionable "Run '...'" hint, matching the project's error convention.
Reject project-mode installs that target the current project root and add a copy-layer guard that skips destination subtrees during recursive copies. This prevents install ./ -p from recursively copying .skillshare/skills into itself on Windows.
A corrupt or unreadable .metadata.json caused LoadMetadataWithMigration
to return a nil store whose error was discarded, so update --all
dereferenced a nil MetadataStore in GetByPath and crashed with a raw
stack trace. Guard GetByPath against a nil receiver and surface the
load failure as a resolve warning instead.
Trim SKILL.md suffixes without filepath.Dir so traversal segments remain visible to SEC003 subdir validation. Add regression coverage for GitHub, GitLab, and Bitbucket blob-style skill URLs.
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.
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.
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
The web UI skill preview always fetched SKILL.md via api.github.com, so skills from GHE or other non-github.com hubs showed only index metadata (name, description, tags) instead of the full rendered content.
FetchPreview now resolves the source platform: github.com and GHE use the host's REST Contents API, while other platforms (GitLab, Bitbucket, Azure, SSH) fall back to a shallow clone. Sources whose content can't be retrieved degrade gracefully with a clear notice in the modal.
Fixes#193
Hub index fetch previously handled only unauthenticated HTTP(S) GET or
local file paths, leaving private and GitHub Enterprise repos out of
reach — raw HTTPS URLs redirect to login before token headers apply, and
SSH-style hub URLs were unrecognized.
Resolve SSH hub sources by shallow-cloning the repo with the user's SSH
keys and reading the index from it. The in-repo index path uses the
existing //path subdir convention (consistent with `install`), defaulting
to skillshare-hub.json at the repo root. Both scp-style and ssh:// URLs
are accepted. In the web dashboard, SSH clones are restricted to saved
hubs so a request-supplied URL can't drive an arbitrary clone.
Ref #185
md2codex / md2gemini named the *input* (markdown) and read poorly. Rename
the bundled extensions after what they produce: codex-agents and
gemini-commands. Turn the BuiltinExtensions catalog into structs carrying
a human-readable description so the dashboard can explain each one before
install, and propagate the new names across tests and docs.
Add an "Extensions" tab to the Config page for managing transform
extensions in the active mode (global or project):
- GET /api/extensions lists installed extensions with their manifest
descriptions, merged with the built-in catalog so users can see what
is available to download.
- POST /api/extensions/install downloads a whitelisted built-in
extension into the active mode's extensions directory.
- POST /api/extensions/open opens that directory in the user's editor,
mirroring the skill "open in editor" affordance.
ExtensionSpec now carries the manifest description so the list can
surface it. The UI shows installed extensions with a built-in badge,
available built-ins with a Download button, and an Open-directory
action. Strings are translated across all 11 locales.
The mixed-tracked-repo recovery path used a fragile substring match on
the error message and re-cloned the repo via /discover to count skills
and agents. Both problems share one root cause: the counts InferTrackedKind
already computed during the failed first install were thrown away.
- Add install.TrackKindAmbiguousError carrying the discovered Skills /
Agents counts. InferTrackedKind returns it on the ambiguous branch.
Error() preserves the legacy message verbatim so CLI output is unchanged.
- Server handleInstall track branch uses errors.As to detect it and emits
writeCodedError with stable code 'install.track_kind_ambiguous' plus
skills/agents counts in error_params.
- UI replaces the substring match with an ApiError.code check and reads
counts straight from err.params, skipping the second /discover clone.
- Add unit tests asserting errors.As, count fidelity, and explicit-kind
short-circuit (no clone).
Net result: 2 clones instead of 3 on the mixed-track recovery path, and a
stable cross-language contract instead of a brittle English substring.
After Phase A (210605ea) fixed metadata + status display for tracked
repos with a root-level SKILL.md, three output gaps remained:
- 'Found 0 skill(s)' on install (should be 1 — root SKILL.md counts)
- 'tracked_repos[].skill_count' was 0 (same root-case)
- 'Run sync' suggestion appeared even when no skill was discovered
Root cause: install/update used discoverSkills(destPath, false) which
walks the tracked repo's own root, treating the root SKILL.md at
relPath='.' as ignored. But 'skillshare sync' walks the source
directory; the same SKILL.md inside _<name>/ has relPath='_<name>'
(not '.') and is discovered as a normal skill. The two walkers used
the same helper with different includeRoot settings and produced
inconsistent counts.
Changes:
- installTrackedRepoImpl and updateTrackedRepo: switch to
discoverSkills(repoPath, true) so SkillCount matches sync's view.
- buildTrackedRepoJSON: when counting skills per repo, accept both
RelPath='<repo>/...' (nested) and RelPath='<repo>' (root-only).
- handleTrackedRepoInstall: suppress the 'Run sync' next-step when
the repo had no skills/agents discovered.
Closes#163
Co-authored-by: Claude <noreply@anthropic.com>
Add azure_hosts config parameter (analogous to gitlab_hosts) that allows
users to register self-hosted Azure DevOps Server hostnames so they
receive proper URL parsing without appending .git to clone URLs.
- Add AzureHosts field to Config and ProjectConfig structs
- Add SKILLSHARE_AZURE_HOSTS env var support with merge semantics
- Generalize hostname validation into shared normalizeHostList/mergeHostsFromEnv
- Add azureOnPremPattern regex and parseAzureOnPrem for /_git/ URL routing
- Add isAzureHost for config-based host detection (no auto-detect)
- Update TrackName to handle on-prem Azure DevOps clone URLs
- Wire AzureHosts through InstallContext, ParseOptions, and Server
- Add JSON Schema, UI fieldDocs, and website documentation
- Add 19 new tests covering config, source parsing, and server
Closes#147
Add PATCH /api/resources/{name}/source endpoint to update the source URL
for installed skills. For tracked repos, all sibling skills sharing the
same git remote are updated together with the new repo base URL, and the
git remote origin is also updated.
Frontend changes:
- Add Source URL input field in the skill editor (visible for all skills
with metadata)
- Use shared Input/Textarea components in FrontmatterEditor for visual
consistency, with new size='sm' variant
- Replace custom kind-badge/tracked-badge with shared KindBadge/SourceBadge
components in the editor header
- Preserve original YAML frontmatter formatting when only the body or
source URL is changed (avoids spurious quote style / line wrap diffs)
- Skip diff review dialog when only source URL is modified
Backend fixes:
- Two-pass skill lookup in handleGetSkill: prefer exact FlatName match
over baseName fallback to prevent name collisions between standalone
and tracked repo skills
- Add basename-only fallback in MetadataStore.GetByPath for tracked repo
skills stored with short keys
- Use cfg.Source (global) consistently in save/patch handlers to match
the source used by list/detail handlers in project mode
Closes#142
- Address review: only RemoveAll discovery temp dirs when Source is non-nil and IsGit(), avoiding deletion for local paths, nil Source, or unknown types.
- Add TestCleanupDiscovery_SkipsWhenSourceNil.
- Local discovery sets RepoPath to the user's directory.
- the UI discover handler defers CleanupDiscovery and was deleting that tree after the response.
- Skip RemoveAll for SourceTypeLocalPath so UI local installs match CLI behavior.
- Adds TestCleanupDiscovery_PreservesLocalSource regression coverage.
Root SKILL.md in multi-skill repos has no directory boundary — the entire
repo root is its source, which drags in project infrastructure (src/,
assets/, CI, build scripts). Now when Path="." and descendants exist,
only the SKILL.md file is copied. Children install as independent flat
directories instead of nesting under the root.
Remove orchestrator nesting logic from installSelectedSkills — each skill
now installs at the top level regardless of whether the root is selected.
Two related improvements for issue #124:
1. Source parsing now tracks ExplicitSkill when a blob URL points
directly at a SKILL.md file. Discovery constrains results to
the root skill only, skipping the orchestrator pack/selection
prompt entirely — the user clearly wants that one skill.
2. Single-select (radio) checklist TUI reworked: selection follows
the cursor automatically, space keeps exactly one item active,
enter confirms the focused item. Pre-selected items set the
initial cursor position. This gives immediate visual feedback
and eliminates the confusing empty-selection edge case.
Refs: #124
InferTrackedKind was called before installTrackedRepoImpl's IsGit()
guard, causing --track with local paths to attempt a clone with an
empty URL instead of returning the correct error message.
Add an early IsGit() check so non-git sources fail with the expected
'--track requires a git repository source' error.
An orchestrator repository containing a root SKILL.md together with
child SKILL.md directories (e.g. iOfficeAI/OfficeCLI) previously
installed as a single bundled directory that mirrored the git clone:
copying the root dragged every child skill dir into it, and children
marked selected were silently skipped as "included in parent" while
their computed dest paths were never written to.
Make each discovered skill a first-class install:
- New copyDirExcluding() skips subdirectories by slash-normalized
relative path, so a skill directory is copied without duplicating
any nested skill that is itself being installed elsewhere.
- installFromDiscoveryImpl computes the descendant skill paths of the
skill currently being copied and passes them as excludes. For the
root of an orchestrator repo this drops every child skill dir; for
a non-root skill it drops any sub-skill nested inside it.
- installSelectedSkills stops skipping children after the root is
installed. The existing dest-path logic already places each child
under <parentName>/<childName>, so the children are written as
siblings to the root install.
After the fix, a multi-skill repo with N skills produces N independent
installed directories discoverable by sync, updatable individually.
Refs: #124
URLs like https://github.com/owner/repo/blob/main/SKILL.md previously
became subdir="SKILL.md" and created an install folder literally named
"SKILL.md". Treat a trailing SKILL.md segment on blob URLs as a pointer
to its containing skill directory:
blob/main/SKILL.md → subdir "", name = repo
blob/main/skills/foo/SKILL.md → subdir "skills/foo", name = foo
Also apply the same rule to GitLab (-/blob/...) and Bitbucket (src/...)
web URLs for consistency.
Refs: #124
Extend install/update with agent-aware --json output and apply the same
audit block-threshold gate skills already use.
Install:
- install --json: imply --force; imply --all only when neither
--skill nor --agent is set (so explicit selectors still work)
- project-mode install now routes through cmdInstallProjectParsed
so --json suppresses UI and emits a pure-JSON payload
- agent installs propagate SourceDir so metadata lands in the
agents root, not the install subdir
- validation error lists --agent / --kind / --branch explicitly
Update agents:
- update agents supports --force, --skip-audit, --audit-threshold
- tracked agent repos + metadata-backed single-file agents now run
the same audit gate as skills; blocks roll back from a temp stage
- JSON output returns the final per-agent status
(updated/skipped/failed/up_to_date/...) by merging check results
with batch update outcomes
Audit:
- auditInstalledAgent shares one helper with auditInstalledSkill
(scan + cleanup callbacks); single-file agent scans now use
audit.ScanFile / ScanFileForProject instead of staging the file
into a temp directory
- UpdateAgentFromDiscovery stages to temp, audits, then renames
atomically (rename already overwrites on POSIX)
Refactors pulled in while reworking these paths:
- extract jsonUISuppressor helper in json_output.go and use it
from install/update/update_agents; collapses 4 copies of the
suppress/restore closure dance and the per-error failJSON block
- drop relPath plumbing from installAgentFromDiscoveryInternal
now that ScanFile doesn't need a staged relative path
- list.go render functions now use theme.ANSI() so NO_COLOR is
respected in plain-text output
Tests:
- add TestInstallAgentFromDiscovery_HighFindingBlocked (unit)
- add TestUpdate_Agents_HighFindingBlockedByThreshold (integration)
- add TestInstall_Project_JSON_Agent_PureJSON,
TestUpdate_Agents_JSON_ReportsFinalStatus
- fix TestList_SKILLSHARE_THEME_Light / TestList_NO_COLOR to
write a config before invoking list
- fix TestTargetList_NoTUI_FallsBackToPlainText assertion to match
the new structured Mode: merge output
- fix TestDoctor_AllGood_PassesAll by pinning SKILLSHARE_THEME=dark
so the no-TTY theme fallback warning doesn't fire
- add list_tui_content_test.go covering light/dark glamour base
- list detail pane now picks LightStyleConfig on light terminals
Docs:
- update install.md, update.md, understand/agents.md for the new
flags, JSON shapes, and agents update workflow
- DiscoveredResource gains RepoRelPath and Targets fields
- AgentKind.Discover populates RepoRelPath via findTrackedRepoRelPath
- AgentCheckResult gains RepoPath and dirty status for tracked repos
- InferTrackedKind auto-detects pure-agent repos for --track install
- TrackedRepoResult reports AgentCount and Agents list
Add local-path resource discovery so install flows can detect agents as well as skills. Preserve nested agent paths and suffixless agent metadata keys during batch install and update, and refresh the UI copy to use resource-aware wording on the install surface.
Production code no longer reads or writes .skillshare-meta.json sidecars:
- InstallAgentFromDiscovery: write to centralized store via WriteMetaToStore
- agent_check.go: read from centralized store passed as parameter
- handler_skills.go: read agents from s.agentsStore
- audit.go: readMetaFileHashes reads from parent .metadata.json
- analyzer_metadata.go: extract findMetaEntry to deduplicate walk logic
- uninstall_agents.go: remove entry from centralized store on uninstall
- batchUpdateAgents: pass store to reinstallAgent, remove dead field
- Replace LoadMetadata+nil guard with LoadMetadataOrNew across 6 files
- Remove MetadataEntry.Name (json:"-", never set in production) and
FullName() method (only used in one test)
- Add LoadMetadataOrNew() helper to replace 8+ instances of the
LoadMetadata + nil guard boilerplate