mirror of
https://github.com/runkids/skillshare.git
synced 2026-11-19 02:17:09 +00:00
4ededca4dd8101f1fa10b2834325f7ed2222e1ea
451
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
|
|
4ededca4dd |
feat(audit): check bundled reference containment (#754)
Required local references can currently escape a skill bundle or fail silently when written as images and explicit code paths. Resolve them within the declared root without reading external targets, retain dangling-link compatibility, and share the findings and rule controls with doctor. Keep metadata and integrity fallback reads within the same root so they cannot reopen skipped symlinks. Co-authored-by: GPT-6.1 Sol <noreply@openai.com> Co-authored-by: Willie <egg358358@gmail.com> |
||
|
|
9b5824806e |
feat(audit): detect whitespace padding that hides instructions (#755)
Long blank spans and repeated filler can move meaningful instructions out of view during skill review. Add four padding signals under the static analyzer with Markdown exemptions and original-source line mapping. Reuse rule overlays so global and project toggles and severity/confidence overrides also govern this file-level detector. Cover threshold boundaries, benign examples, optimized scans, reports and CLI rule controls. Implements item 17 of #746. Co-authored-by: GPT-6.1 Sol <noreply@openai.com> Co-authored-by: Willie <egg358358@gmail.com> |
||
|
|
53a3207a17 |
fix(update): keep an agent's source when updating it (#778)
* fix(update): keep an agent's source when updating it After the first `update agents`, an agent's recorded source became `/agents/<name>.md` with type `unknown-subdir` and no repo URL, so `check agents` stopped looking upstream and later changes never arrived (#777). The batch path rebuilt the repository source from the clone URL, subdir and branch only, leaving Raw and Type empty, and the metadata written back was built from that empty Raw. The fallback reinstall and the dashboard update kept the agent's full source as Raw while clearing the subdir, so the written source repeated the agent path. AgentRepoSource now rebuilds the repository source once, by cutting the agent path off the recorded source, and all three paths use it, so the source written back is the one install recorded. Agents whose metadata was already rewritten have no repository left in it and need to be installed again. * fix(check): compare each agent at its own branch and report unchecked agents Agent update checks gave wrong answers in several cases next to #777: - Agents from one repository were compared with the first agent's version, so an older agent sharing a repo with a current one never showed its update. - The installed branch was never sent to the remote, so an agent pinned to a non-default branch was compared with the default branch head. - A tracked agent repo records its full commit while the remote hash is abbreviated, so a fresh clone always read as update_available; its checked-out branch was also ignored. - `update agents` printed "All agents are up to date" and exited 0 when agents could not be checked at all: an unreachable remote, unreadable metadata, or a source erased by the #777 bug. Each agent is now compared with its own version at its own branch, and an agent whose source the old bug erased reports an error asking for a reinstall, since its metadata no longer names a repository. Update lists the agents it could not check and exits non-zero, counting them apart from agents that failed to update. The docs describe the error status and the reinstall fix. * fix(update): repair dashboard-doubled agent sources and share the update flow Review of the previous commits found gaps in recovering agents that earlier updates broke, and in how the two update handlers report them: - A dashboard update before the #777 fix appended the agent path to its source a second time. Those agents keep their repo URL, so check offers an update, but discovery then failed on the doubled path. Every copy of the path is now cut, and the update writes the source back correctly. - On the dashboard, an agent whose source the CLI bug erased failed with "invalid source: source cannot be empty". AgentRepoSource now returns the same reinstall error check shows, and both use one detector. - When every tracked agent was such an erased agent, no remote was asked, so the "Checking agents" spinner was never stopped and drew over the error rows. It is now stopped once the check returns. - The global and project handlers repeated the check-and-update flow, and the project copy had already drifted (no dry-run log entry, spinner in --json). Both now run one shared flow. Tests cover the erased agent next to an updatable one in global, --json, --dry-run and project mode, and the doubled and erased sources in AgentRepoSource. |
||
|
|
11add19544 |
feat(ui): note how Nix updates the CLI in the update dialog (#771)
* feat(ui): show the install's own upgrade command in the update dialog The update dialog always offered `skillshare upgrade` for the terminal. On Nix that command only prints advice, and Homebrew users are better served by `brew upgrade skillshare`, so the right command appeared only after a click on Update returned an error. /api/version now detects the install method, reports its command as upgradeCommand, and looks up the latest version the way the CLI update notice does for that method. The dialog shows that command, or the existing "update your Nix configuration and rebuild" text when a Nix configuration provides skillshare, so no new translations are needed. * fix(ui): keep `skillshare upgrade` in the update dialog, add only a Nix note Review of the previous commit found three regressions: - `skillshare upgrade` already runs `brew upgrade` on Homebrew and also updates the built-in skill. Showing `brew upgrade skillshare` (or the nix profile command) instead broke the case where only the skill is outdated: the shown command updates nothing. - /api/version looked up the latest version through `brew info` on Homebrew, inside the request with no timeout. Homebrew's auto-update could stall the dialog, and a server without brew on PATH never showed an update. - The rendered dialog tests only checked shown text, which wiki/frontend.md says to leave to screenshots. The dialog keeps `skillshare upgrade` and the GitHub lookup. /api/version reports nixUpgradeCommand only for Nix installs, and when a newer CLI is pending the dialog adds the existing Nix advice under the command. The decision lives in nixUpdateNote, unit-tested in upgradeError.test.ts. |
||
|
|
4fd4058be2 |
test(upgrade): cover the Nix branches of upgrade and doctor (#768)
#764 changed how upgrade and doctor treat Nix-store binaries but only tested the version and server layers. A later edit to upgradeCLIBinary could start replacing a store binary, or drop the note the dashboard matches for its 409, without any test failing. Two package variables let the tests fake the executable path and the latest release. The ui.md pages said /api/upgrade refuses Nix installs. It still updates the built-in skill and answers 409 only when a newer CLI is waiting, and the dialog shows `skillshare upgrade` until clicked. |
||
|
|
64c8a3f34e |
feat: add Nix flake and handle Nix-store installs in upgrade (#764)
* feat: add Nix flake and stop self-upgrade of Nix-store installs Add flake.nix with packages, overlays.default, checks and a formatter, building nix/package.nix with the unit tests in the Nix sandbox. The version comes from the Release Please manifest, so releases need no Nix edit. A Nix workflow runs `nix flake check` on non-docs changes. `skillshare upgrade` now detects binaries under /nix/store/ and prints `nix profile upgrade skillshare` instead of replacing a read-only store path; update notices and doctor suggest the same command. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(upgrade): give Nix installs the right update path everywhere The first flake commit sent every /nix/store binary to `nix profile upgrade`, failed a plain `skillshare upgrade` with exit 1 even when the CLI was current, and left the dashboard with a generic 500. - Tell `nix profile` installs (the profile manifest lists the store path) from NixOS, Home Manager and overlay installs, which are told to update their configuration and rebuild. Detection lives in one helper shared by the update notice, doctor, upgrade and the dashboard. - `upgrade` on Nix reports "already up to date" or prints the hint and exits 0; the built-in skill still updates. - The dashboard runs the upgrade so the skill updates, then answers 409 `upgrade.nix` with the command when a newer CLI waits on Nix. The UI shows localized text for it in all 11 locales. - Installs skip the unit tests; the flake's checks enable them. postPatch lowers only the go directive's patch version and fails evaluation with a clear message when go.mod needs a newer Go minor. CI also builds the package on macOS. - Docs, translations, the built-in skill reference and the release topic describe the new behaviour. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
|
8153025beb |
fix(plugins): reconcile OMP source replacements in one sync (#758)
A returning machine can retain an old runtime link after shared plugin intent changes source. Review its verified scoped cleanup together with replacement, then remove the old link before installing the new source while retaining shared cache and settings. Preserve the original recipe identity when cleanup of a pending-only installation fails, so retries still reconcile the replacement. Add filesystem-backed regressions, Hegel ownership properties and a two-machine native CLI runbook. |
||
|
|
d41188bbde |
fix(project): gitignore install metadata in project mode (#760)
Since v0.19.12 the docs say a project's skills/.metadata.json is gitignored, but no code ever wrote that rule, so `git add .skillshare` picked it up (#759). The file records this machine's installs (hashes, timestamps) for folders that are themselves ignored; teammates already share config.yaml and skills.lock.json, so committing it only adds noise and merge conflicts. ReconcileProjectSkills runs after every project install (CLI, search, dashboard, move, trash), so it now adds the rule for the skills and agents sources, at the path ProjectGitignoreTarget computes. Projects initialized earlier and custom sources get it on their next install. AdoptMovedProjectSkills is untouched, so `sync -p` still leaves .gitignore alone (#547). Global mode keeps committing its metadata. The docs that list the project .gitignore contents now name the metadata files, and project-skills explains the one-time `git rm --cached` for projects that already committed it. Closes #759 |
||
|
|
65af9c17b7 |
feat(audit): detect narrow prose risk directives (#711)
* feat(audit): detect narrow prose risk directives * fix(audit): honor link confidence and narrow refusal rule Apply configured confidence to Markdown link findings and require safety-related evidence in anti-refusal directives to avoid ordinary compliance false positives. * fix(audit): address prose and confidence review feedback Narrow prose matches, preserve overrides when enabling rules, and expose effective confidence to rule consumers. Expand Markdown coverage and synchronize the audit guidance. * fix(audit): retain scheduled payload detection Keep detecting directives that schedule a payload while excluding ordinary payload processor terminology. * fix(audit): preserve prose detection recall Recognize concise disclosure and safety override directives while retaining benign exclusions. Share the confidence default and cover replacement overlays. * fix(audit): detect quantified refusal directives Retain detection when a directive refers to any or every user request while preserving benign security discussion. |
||
|
|
c6ec6bc7c9 |
fix(audit): distinguish execution errors from findings (#686)
* fix(audit): distinguish execution errors from findings Return a separate exit code for audit errors so CI can tell an incomplete scan from a threshold finding. Preserve partial reports and give scan errors precedence when both occur. * fix(audit): keep list and abort exit codes consistent Preserve audit execution-error classification when list delegates to audit, so both global and project actions exit 2. Treat declining a large scan as a clean cancellation before any scanning or error logging. Cover both paths with real terminal regressions and keep the documented exit-code contract aligned across translations. |
||
|
|
51d299ff07 |
fix(check): ignore OS metadata files when detecting local changes (#744)
Finder writes .DS_Store into any folder it opens, and Explorer writes Thumbs.db and desktop.ini. These counted as skill content, so opening an installed skill in a file manager made check report it as modified, hid update_available, and made update skip it unless run with --force, which also lowers the audit threshold. Skip them when hashing a skill, drop hashes earlier versions recorded for them (from .metadata.json and from legacy sidecar migration), and keep audit integrity from reporting them. Their content is still audited. Closes #743 |
||
|
|
e3ab0cbeff |
docs(mcp): describe the Agent logos on enabled global server rows
The project MCP tab now shows which Agents load an enabled global server
(
|
||
|
|
78a8dc7c0d |
fix(trash): keep install metadata when restoring from the trash (#680)
* fix(trash): keep install metadata when restoring from the trash Uninstall pruned a skill's, tracked repo's or agent's entry from the central .metadata.json, and only the files went to the trash, so a restore brought back a remote install as a local one: update and check no longer knew its source (#679). Uninstall now saves the entries it prunes, with their audit acceptances and target overrides, beside the trash entry, re-keyed to the trash name so they match where restore puts the item. uninstall.Restore and RestoreAgent merge them back after the files return, and every restore path (CLI, trash TUI, dashboard) goes through them; the dashboard passes its cached store so it does not go stale. The saved copy is taken with the same rule RemoveByNames uses and copies each entry, so legacy repo-group entries are kept and re-keying never touches the live store. In project mode uninstall also drops the skill's .gitignore rule and lock pin, so a skill restore reconciles the project skills afterwards. trash.Delete removes an entry with its saved metadata, and replaces the direct RemoveAll calls in delete, empty and the 7-day cleanup. Items trashed before this change have no saved metadata and still restore as local. * fix(trash): keep live records of another path when restoring metadata A legacy store keys a grouped skill by its basename, so the saved copy of frontend/foo was keyed "foo". If a different skill was installed as root-level foo while frontend/foo sat in the trash, restoring it replaced the root skill's record and lost its source. Extract now writes full source-relative keys (legacy repo-group entries keep theirs, since their Group is not their path), and Merge keeps a live entry whose key names a different path instead of overwriting it. Merge is shared by skill and agent restore, so both are covered. |
||
|
|
00abba4727 |
feat(mcp): allow plain HTTP on trusted private networks (#675)
* feat(mcp): allow plain HTTP on trusted private networks Containerized MCP gateways need to reach a skills server directly on a private network without adding a TLS terminator. Add an explicit flag and matching serve config opt-in, retain bearer authentication and the secure default, and warn when plaintext transport is enabled. * fix(mcp): require explicit HTTP transport opt-in Keep insecure HTTP opt-in in the container command so repository configuration cannot enable it. Reject the flag without HTTP, align translated examples, and cover startup warnings and authentication in the runbook and transport tests. |
||
|
|
da3383ed5f |
fix(plugins): separate shared intent from local installation state (#660)
* fix(plugins): separate shared intent from machine-local receipts Shared configs currently carry native paths, reviewed snapshots and pending failures, so one host can publish local progress over another host's target selection. Persist versioned declarations separately from local applied and pending receipts, and reconcile reviewed installs, imports and cleanup against this machine's native inventory without rewriting intent during sync/update. Keep snapshot names and adapter safeguards, move the operation lock locally, validate legacy adoption with local evidence, and invalidate approval on local state or native-context changes. Preserve source recipes after native adoption, require local import for absent native-only targets, and document the shared writer upgrade boundary in schemas and translated references. Cover two-host behavior, generated legacy fixtures, failure/recovery, stale intent, cleanup, npm recipes and the existing CLI/API/UI consumers with focused resource-capped tests. Fixes #651. * fix(plugins): preserve intent and ownership across host transitions Repeat-add must retain shared target selection, and update previews must resolve the requested ref before reviewing fresh or replacement installs. Keep disabled targets on the uninstall path and reject refs unsupported by native-only/npm intent before they can produce invalid shared declarations. Block changed-context add/import before Pi registration preparation so a new account cannot overwrite the old receipt. Preserve unresolved legacy bindings as passive private proof candidates before rewriting shared YAML, retaining original intent/store for later verified adoption and cleanup. Never compose these hints as executable state, and persist their explicit resolution. Cover the five review classes, native/private failure boundaries, requested-ref retry, and original-context hint proofs with isolated focused regressions. Keep local format v1 and baseline account validation unchanged. Refs #651. * fix(plugins): allow receipt-only orphan cleanup to complete An installation already removed on this host can retain a local receipt when another host removes or replaces shared intent. Sync correctly plans forget, but native command verification rejects that metadata-only action and blocks all targets before the receipt can be cleared. Exclude forget from native command verification while retaining host inventory checks and OMP scoped cleanup preparation. The existing completion path then persists receipt deletion, including pending work and passive legacy hints. Cover the cross-host disable/re-enable/remove walk, five native adapters, failed work with a legacy hint, and absent source replacement alongside a healthy target. Assert Apply completion, unchanged shared bytes/mtime, and a quiet second sync. Refs #651. * fix(plugin): reconcile ref intent across inventory consumers Validate explicit ref changes before disabled targets can save them and reuse one reviewed discovery throughout a preview. Surface shared ref changes as local reconciliation work while preserving the applied native identity and receipts. Reuse scoped legacy proofs for Pi extension ownership and make full inventory the dashboard's authoritative package/host pair. Cover invalid and replacement refs, read-only ownership, and shared-state properties without new persisted state or dependencies. * test(plugin): repair and consolidate host-state fixtures Create source intent and a pending imported Claude update through supported operations so direct and batch updates exercise the intended fixture. Keep the marketplace-loss case locally owned before testing its skipped reinstall. Consolidate cross-host repeat-add and requested-ref replacement/orphan retries at their existing owners, retaining failure recovery and shared-byte checks. Trim duplicate HTTP and translated-text assertions from consumer tests. * fix(plugin): reconcile legacy snapshot reads and native imports Resolve legacy snapshot ownership through existing scoped proofs before listing, reading or discovering cached source files. Keep adoption read-only and limit proof work to the requested package and targets; current receipts need no CLI. Only preserve reviewed source metadata on reimport when intent remains source managed. Native-only intent must not resurrect an older Git source after a same-ID local import. Retain identity mismatch and reviewed-source safeguards. Remove the unused source-folding helper left by the old shared runtime writer. * Preserve reviewed package versions locally * Fix reviewed package version lifecycle * fix(plugin): report missing local imports in doctor Show the local-import instruction instead of claiming a skipped imported plugin is healthy. Preserve the missing-CLI warning and shared config bytes. Keep fixture native settings inside scratch XDG directories and serialize the fake CLI calls used by concurrent inventories. Update assertions to separate shared declarations from local versions and native IDs, and align the raw-write audit with private plugin state and operation locks. * test(plugin): use declarations for doctor source fixture The missing-CLI case has no native ID, so encode it with the version 2 declaration schema. Keep the native-only import case in legacy syntax. * fix(plugin): install shared sources from reviewed snapshots When a source target is already declared, first sync must install its reviewed local snapshot even if the native client has a package from the same marketplace. Keep explicit add adoption and native-only import intact. Extend the existing two-host adoption check to cover an independent native installation on the second host and verify a quiet subsequent sync. * fix(plugin): preserve reviewed retries and native collision boundaries Retry a pending source update through the existing reviewed update path, so a rolled-back native failure cannot clear progress at the old snapshot. Record default-source versions for add, sync, and update while retaining the recipe fingerprint and generic-version checks for target overrides. Keep explicit native adoption, but block shared Claude sources when an independent same-name native plugin could shadow their hooks. Extract the existing adoption decision so the preview function becomes smaller. * test(plugin): respect OMP update safety in version control Check generic and target versions during review, then confirm the existing OMP upgrade guard preserves the installed and default-source versions. * fix(plugin): declare first local native imports explicitly A first local import into a shared source target now changes that target to native-only intent while preserving its sync selection. Verified source reimports retain their reviewed source and the preview discloses conversion. Cover the explicit import, local sync, and other-host cleanup flow, and remove the unconsumed internal observation assertion from the OMP test. * docs(plugin): explain first local import intent changes Describe the shared native-only target conversion and its effect on other machines, while distinguishing verified source reimports. * fix(plugin): verify checked sources before native reinstalls A checked update that becomes an install after native removal must still match the source digest the user approved. Reject moved or unchecked sources before native work, and allow a stable approved reinstall. Clear completed reinstalls from dashboard update availability while keeping failed targets available for retry. Cover both native presence states in the existing checked-update tests and the dashboard outcome rule. * fix(plugin): guard replacement before cleanup * fix(plugin): refresh discovery for each manager action A later Add action must review the current source even when the manager previously observed that source. Keep discovery sharing across one action while ending its cache before the next wizard invocation. Exercise a fresh source version after an earlier Add preview without TTY input or native writes, using an unavailable client to stop at review. * test(plugin): cover desired checks and final update outcomes Consolidate checked batch approval and dashboard completion tests while retaining removed-native, digest, version and approval controls. Reproduce stale shared recipes and failed or unfinished replacement installs. * test(plugin): make stale source discovery observable Remove the old repository before native-only and npm desired checks so those existing cases cannot pass by reading the applied source. * fix(plugin): check desired source and final outcome * test(plugin): bind checked updates to the observed recipe Reproduce a shared candidate change with equal version, commit and digest, and preserve a freshly checked replacement whose native ID differs. Use actual check observations in approval and native update controls, reject missing source identity, and verify the UI carries the observation through. * fix(plugin): bind checked update to source recipe * fix(ui): classify npm recipes as Pi packages * fix(plugin): retain legacy removal intent during migration Omit pending removals from shared declarations while retaining original local ownership proof so sync retries cleanup. Verification: focused regression passed in an isolated Linux Codex container, not the devcontainer (GOMAXPROCS=4, -p=4). * fix(plugin): keep imports local to an existing shared recipe Adopt a native installation into its local receipt without converting the shared source target or removing other machines installations. Verification: focused import regressions passed in an isolated Linux Codex container, not the devcontainer (GOMAXPROCS=4, -p=4). * fix(plugin): capture legacy evidence on first read Persist passive hints before another machine writes v2 and save verified adoption on reads. Apply reuses preview receipts instead of querying every legacy binding again. Document the per-machine pre-migration read. Verification: migration regressions passed in isolated Linux Codex container runs, not the devcontainer (GOMAXPROCS=4, -p=4). * fix(plugin): target named declarations in interactive imports Offer unresolved package names in dashboard and TUI imports and pass the selected name in single and batch requests. Drop display-only rendered assertions. Verification: 47 UI tests passed with at most four workers in an isolated Linux Codex container, not the devcontainer. * fix(plugin): localize shared intent plan guidance Add message keys for local imports, account changes, cleanup and native collisions across all dashboard locales. Localized guidance contains no CLI flags. Verification: failing localization regression now passes in an isolated Linux Codex container, not the devcontainer (GOMAXPROCS=4, -p=4). * refactor(plugin): reuse preview inventory and guard apply boundaries Capture hints on first load without persisting in-memory adoption during preview, keeping approval revisions stable. Apply reuses preview inventory. Reuse accountEnv and guard applyChange plus native write boundaries without inspecting argv. Verification: migration, receipt, account and sync regressions passed in an isolated Linux Codex container, not the devcontainer (GOMAXPROCS=4, -p=4). * test(plugin): organize host state coverage around lifecycle scenarios Group migration, ref drift, account context and cleanup tests behind shared host fixtures and public Preview/Apply assertions. Cover Codex, Cursor, Antigravity and project scopes; add second-machine doctor coverage and the two-machine runbook. Consolidate package-version tests and remove the unused private-record fixture. Verification: focused plugin regressions passed in an isolated Linux Codex container, not the devcontainer (GOMAXPROCS=4, -p=4). * docs(plugin): explain local imports and per-machine migration Document local receipts, per-machine review and commit pinning, config-path state identity and the required legacy read on every machine. Align sharing, doctor, troubleshooting, file-layout and four translated plugin references; restore schema editing guidance. Verification: context-router check passed. Runtime verification uses an isolated Linux Codex container, not the devcontainer. * refactor(plugin): separate unrelated review scope Move project OMP configuration, plugin list ordering, and npm Pi classification to independent local branches based on upstream/main. Keep the shared intent change focused and cover named import selection in the pure flow helper. The three local branches have passing builds in a Linux Codex container, not the devcontainer, with isolated test homes and capped parallelism. * perf(plugin): avoid recapturing v2 state on reads Only compare and persist passive migration hints when legacy bindings exist. Native v2 reads need no state serialization for migration. Verification: pending removal and first-read/import regressions passed in a Linux Codex container, not the devcontainer, with isolated HOME/XDG/TMPDIR and GOMAXPROCS=4 GOFLAGS=-p=4. * docs(plugin): clarify native marketplace repair after import A local import receipt is not a reviewed reinstall source. Correct the remaining troubleshooting instruction and its four translations to require native-client repair, and align the nearby code comment. Validation: checked against prepareChange, context-router check and git diff --check passed. Verification for this review ran in a Linux Codex container, not the devcontainer. * fix(plugins): preserve local ownership across shared intent changes Retain completed legacy migration markers so unchanged v1 YAML cannot uninstall a later manual reinstall. Serialize read-side hint persistence with native operations and keep local imports when shared recipes change. Require explicit dashboard import names and validate recipe identities. Correct schema descriptions, messages, CLI ref status, translated migration guidance and runbooks. Add failing-first Preview/Apply regressions and a real two-ssenv CLI lifecycle script. * fix(plugins): preserve migration history and show retained import guidance Mark existing legacy hints and receipts as captured before cleanup so state from the previous serializer cannot re-adopt a later manual installation. Keep recipe-change guidance in unchanged sync outcomes and visible dashboard rows, and resolve nested OpenCode entries through their package manifest. Cover public migration, ref, account, import and message contracts across adapters. Correct the real two-host Cursor assertions, explicit dashboard selection flow, and remaining localized guidance. * fix(plugins): retain explicit OpenCode entry during import matching Standalone OpenCode packages can rely on a reviewed explicit entry rather than an SDK dependency. Preserve that entry when reading native package identity so matching nested imports remain supported. * fix(plugins): preserve migrated imports and verify native entry identity Recognize older native-only receipts by their existing declaration fingerprint so remote source recipe edits cannot replace migrated imports. Keep managed npm receipts distinct through their source fingerprints. Only borrow an ancestor OpenCode package identity when its resolved entry is the actual registered file. Retain nested and explicit entry support. Add public two-host migration and mismatching-entry regressions. * fix(plugins): retain imported ownership through repeated adds Persist inferred imported ownership when a matching add refreshes the receipt intent, so later shared recipe changes still preserve local native installations. Keep the legacy replacement regression aligned with that ownership contract while retaining removed-intent cleanup. Reject missing or unstatable absolute OpenCode entries before identity matching rather than falling back to the file name. Cover both paths through public Preview and Apply regressions. * test(plugins): isolate OpenCode adapter fixtures from host config Keep native registration fixtures inside their fake agent home by setting XDG_CONFIG_HOME and clearing inherited OpenCode overrides. Ambient config writes can otherwise leak into config-discovery tests in other packages. Cover the XDG, file, directory and inline override cases with temporary config sentinels, while retaining the existing MCP listing expectation. * fix(plugin): let local imports return to the shared source Add a local unimport action that removes only this machine’s installation and receipt while preserving the shared source declaration. Keep retained-import guidance visible across matching adds, list, doctor and dashboard consumers. Avoid idle OMP guidance and unchanged legacy-read locks, preserve localized errors, keep import identity checks offline, and normalize OpenCode drive URLs. Cover the A/B transition and align translations, runbooks and environment cleanup. * fix(plugin): complete unimport validation and guidance Reject missing targets before planning local unimports and share per-target eligibility with the wizard. Make native uninstall effects explicit in help and translated docs, preserve structured update guidance, and cover error paths and shell completion. |
||
|
|
7ec788dea4 |
feat(mcp): accept agy and antigravity-cli as MCP targets (#670)
The agy CLI keeps its MCP servers in the file the antigravity client writes, and plugins and hooks already call it antigravity-cli, but mcp --target and --from rejected that name. Map agy and antigravity-cli to antigravity when the flags are parsed, so the source still stores one client name. Document the aliases in the MCP command reference, its four translations and the built-in skill reference. Closes #669 |
||
|
|
85bcf589a7 |
feat(sync): recognize and route agent-specific skill variants (#649)
* docs(proposals): propose agent-specific skill variants Tracked repositories can ship different files for the same skill across agents. Counting and syncing each copy independently sends incompatible instructions to every target unless users maintain routing filters. Propose repository-scoped logical identity and exact variant routing while preserving source files, explicit controls, and existing naming behavior. Leave activation and compatibility decisions for maintainer review. Refs #648 * docs(proposals): clarify variant routing compatibility Explain how recognition affects universal-only targets and how existing per-variant overrides preserve the desired routing. Fix the precedence between variant identity and generic host-style path inference. Include analyze, move, and catalog consumers while preserving physical identifiers and actual context-cost accounting. Refs #648 * docs(proposals): clarify activation assumptions and scope Make acceptance cases conditional on maintainer-approved recognition being enabled for the tracked source. Separate the provisional implementation estimate from this proposal-only PR so its size is clear. Refs #648 * docs(proposals): qualify naming and opt-in assumptions State that the reproduced routing behavior uses default flat target naming. Leave any opt-in surface with the maintainer's activation decision without ruling out future CLI or config fields. Refs #648 * feat: recognize and route tracked agent skill variants Count tracked variants as logical skills in source and repository summaries, while preserving physical paths, sync operations, and context costs. Select matching canonical agent variants, including account targets, through shared discovery and selection with existing explicit routing precedence. Refs #648 * fix: count logical skills in sync inventory summaries Use shared logical and physical inventory counts in global and project sync closing lines while preserving each target's operation totals. Refs #648 * fix: preserve variant routing diagnostics across previews Resolve target selections once so CLI and server previews keep ambiguity reasons and physical skip counts. Explain managed variants removed by automatic agent routing, and report logical inventory in doctor. Keep warning rows read-only, document upgrade recovery in translated filtering references, and remove the optional README-only addition. Refs #648 * fix: align sync matrix previews with shared variant selection Classify the full source selection once per target so matrix previews honor canonical aliases, account identity, exact-agent routing, and ambiguous variant skips. Preserve each rejected physical path's existing diagnostic and draft filter disambiguation; skipped preview rows remain read-only. Refs #648 * test: expect singular skill wording in doctor source checks Keep the undiscovered-link count assertion aligned with logical source inventory formatting for one skill in global and project mode. Refs #648 * fix: retain selection diagnostics in catalog and diff verdicts Report warning-only diff targets as requiring attention while preserving local-only folder semantics. Reuse shared per-path selection diagnostics in target-scoped catalogs without applying target naming to served paths. Compute source counts once in summary callers and require variant counts in the source status formatter. Refs #648 * test: tighten tracked variant preview fixtures Use a real committed tracked repository for status counts, verify sync preview banners retain server warnings, and use emitted ambiguity reason text in target preview fixtures. Refs #648 * fix(sync): keep agent variants in shared target folders Compatible targets sharing a skills folder need a combined selection so Codex and universal do not prune one another in merge or copy mode. Carry every shared reader through selection, pruning and previews, including project targets, account identities and path aliases. Require a matching skill suffix under two distinct agents before inferring variants, keeping ordinary agent-named categories unrestricted. Separate naming warnings from ambiguous selection in diff and explain ambiguous removals. Add global/project regression coverage for repeated sync, target-only sync, status JSON, diff and API consumers. * fix(variants): align consumer counts, hints and diagnostics Use logical source counts in list, tracked install and init summaries while matching audit caches against physical variants. Restore localized dashboard skip hints and explain symlink routing limits in status and doctor, including project and JSON output. Document paired-agent detection and compatible shared-folder selection in the proposal, built-in skill and English plus translated command pages. Add focused regressions for the affected consumers. * fix(sync): cover initial sync and ignored variant diagnostics Init syncs a fresh in-memory config, so resolve shared target identities before its first sync as well as when loading saved configuration. A whole-directory symlink exposes ignored skills too. Diagnose agent variants using physical inventory when active discovery has none, while keeping logical counts and merge/copy selection restricted to enabled skills. Cover global/project JSON status and doctor with regressions. Reuse sourcewalk directory and canonical-path helpers so traversal guards remain enforced and path resolution has a single existing owner. * fix(variants): address conversion, counts and selection diagnostics Ignore runtime agent identities when comparing convertible project targets and count logical skills during init. Keep selection warnings separate from naming warnings and preserve agent-named parent skills as ordinary skills. Cover account sync/prune, project ambiguity and overrides in merge/copy modes, and init/status agreement. Focused verification ran in an isolated Linux Codex container, not the devcontainer, as authorized. * refactor(variants): remove obsolete selection helpers Remove the unused FilterSkillsByTarget wrapper and prune agent parameter, cache canonical target aliases once, and restore catalog skip ordering to avoid an unrelated behavior change. Completes the optional cleanup from PR 649 review. * fix(ui): expose localized variant selection diagnostics Show target warnings with source paths, localize ambiguity across preview and detail consumers, and let skipped preview rows add a physical exclude pattern. Show tracked repository variant counts alongside logical skills. Drop display-only rendered cases while retaining mutation protection and plain TypeScript logic coverage. Inspect target states in Clean/Playful light/dark and Traditional Chinese. * docs(variants): explain diagnostics and add sync runbook Document doctor variant_routing and logical/physical counts, diff selection skips, and the nested parent-skill boundary. Keep all four command translations aligned and add isolated account/project E2E scenarios. Both runbook steps passed in the authorized Linux Codex container with separate HOME/XDG/TMPDIR fixtures. Website typecheck and all five locale builds passed. * test(config): use native paths in project conversion fixture The shared-agent regression fixture appended slash-separated suffixes to Windows temporary paths, so project path matching returned no candidates. Build each path with filepath.Join and preserve all conversion assertions. Focused project configuration tests passed in the authorized isolated Linux Codex container, not the devcontainer. Windows CI verifies the platform-specific fixture correction. * fix(ui): recompute shared skill routing for filter previews Evaluate edited include/exclude settings in the complete target map before classifying the draft, so a formerly shared folder no longer keeps its old union when the filters differ. Preserve the saved target configuration. * fix(mcp): keep parent skills when peer variants route elsewhere Do not treat an otherwise eligible variant for another agent as an unserved nested skill. Continue blocking parent publication for explicitly filtered, disabled, ambiguous, or invalid children so filters still hide their files. * fix(ui): distinguish agent selection from naming skips Expose separate naming and selection skip counts on target summaries. Show each applicable remedy in target details and sync previews so ambiguity keeps its filter/override guidance alongside naming and collision hints. * fix(list): include disabled rows in tracked repository counts Count the complete inventory that list displays, including ignored skills and physical variants, without changing the enabled counts used by status, install, or sync. Clone rows before counting to preserve disabled state. * docs: align variant counts and routing recovery guidance Show logical skill and physical variant counts in list/install documentation and keep the four translations aligned with the filtering examples. Explain that agent-named skill parents keep ordinary nested skills and retain the devcontainer-only runbook environment. Removal warnings now recommend a source-local override, since include filters cannot undo agent routing. * fix(mcp): preserve explicit target restrictions on nested variants Limit the differently routed child exemption to automatic layout inference. Frontmatter targets and dashboard source-local overrides must still prevent a parent from publishing files explicitly restricted to another target. * fix(ui): fill target name in naming warning guidance Pass the current target name when rendering localized warning hints so the same naming guidance shown beside the controls does not expose a placeholder in the warning banner. Verified in the populated target preview. --------- Co-authored-by: Willie <egg358358@gmail.com> |
||
|
|
23c9453f87 |
fix(update): serialize checkout updates across skillshare processes (#666)
Two dashboards (or a dashboard and the CLI) serving the same source ran "Update all" at once. The server mutex only covers one process, so both pulled into the same checkouts: one saw the other's half-applied merge as "local changes would be overwritten", and concurrent fetches into a blob:none clone failed with "fetch-pack: invalid index-pack output". update.TrackedRepo, install --track --update, and the in-place pull of a git-installed skill now hold a flock in the checkout's git directory; a second process waits up to a minute, then fails with a clear message. Force on a git-installed skill only skipped the local-modification check and then ran a plain pull, which refused to merge over the edits the docs say --force discards. It now fetches and resets to upstream, so a failed fetch still keeps the edits. Followed checkouts keep pulling, since they must be clean and may hold the user's unpushed commits. The pull error dropped the file list git indents under it, and the finished-update subtitle said everything was up to date even when items had failed. |
||
|
|
51fa8d14b3 |
fix(sync): summarize missing registered project roots (#665)
Shared configs can register roots that exist on other machines. Replace the per-root warning flood with one count and a hint to project list, which already exposes each missing folder in text and JSON. Keep root skipping and sync JSON unchanged. |
||
|
|
89ae7068d5 |
fix(config): honor OMP agent directory for default skills path (#652)
* fix(config): honor OMP skills directory overrides OMP reads global skills below PI_CODING_AGENT_DIR, but unpinned built-in targets still selected the default home. Resolve that runtime default without changing the canonical template used to derive named accounts. Keep account scan paths independent of the override and document explicit path precedence in target guidance and its existing translations. * fix(config): preserve inferred OMP paths across saves Saving an unpinned OMP target serialized its resolved skills path, so a later agent-directory override no longer selected the native location. Track this fallback with the existing derived-path marker so config saves preserve its inferred status while explicit paths remain pinned. Cover save/reload after environment changes and filter edits, and fold redundant account and project-alias checks into their existing cases. * docs(config): clarify when OMP skills paths stay fixed Initialization and target addition save their selected destination, while an OMP entry with no configured skills path follows later environment changes. Clarify that distinction in public and built-in target guidance and all four existing translations. Keep project alias coverage in its existing owner test and remove redundant OMP matcher and generic save assertions. * test(config): normalize explicit OMP account paths The explicit-path fixture combines a native temporary root with a forward slash suffix. Clean the actual path for comparison so mixed separators match the native expected path on Windows, retaining the original path in failure output. * test(config): use portable OMP whitespace override fixture Windows normalizes a directory component ending in a space during native absolute-path resolution. Use a leading-space component to preserve the check that override values are not trimmed without depending on trailing space filename behavior. * fix(config): preserve default OMP home in account shells Keep plain OMP skills on the canonical home when its environment override matches a declared account, so the unfiltered default target cannot bypass account filters. Share account-aware defaults with discovery and dashboard selection, and keep canonical account templates separate from runtime homes. Require an OMP install before automatic detection because Pi shares the same override. Trim surrounding whitespace, ignore relative overrides, and retain explicit saved paths. Cover account filtering, aliases, symlinks, saves, discovery, and existing environment-selected file homes with regressions. |
||
|
|
eb9b2b28f9 |
feat(plugin): add --all flag to plugin update and enable antigravity-cli updates (#656)
* Add --all flag to plugin update and enable agy Add -a/--all to skillshare plugin update so users can update all tracked plugins in a single command without tying slow target CLI runs to skillshare update --all or sync --all. - Reject --all outside plugin update or when combined with a plugin name or --source-ref - Share source repository discoveries and discovery errors across the Preview and Apply lifecycle via Service.WithDiscoveryCache so multi- plugin marketplaces are cloned at most once per command - Isolate source, manifest, and target failures per binding when --all is active so healthy plugins update while partial failures exit non-zero and leave failed bindings unchanged in config.yaml - Skip unfiltered OMP bindings during --all updates so sibling targets on the same package still update without failing the batch - Enable plugin update for antigravity-cli while skipping disabled plugins when agy extensions list reports an entry as disabled Closes #653 * Add plugin update --all to docs and completions Advertise -a/--all for skillshare plugin across all five shell completion scripts and document the bulk update workflow and antigravity-cli update support in the user and built-in skill docs. - Add --all and -a to plugin completions in bash, zsh, fish, PowerShell, and Nushell with TestPluginCompletionsIncludeAllFlag - Document update [NAME | --all], -a/--all, and antigravity-cli source updates in website/docs/reference/commands/plugin.md (and its zh-Hans, zh-Hant, ja, and ko translations), sharing-plugins.md, and quick-reference.md - Document plugin update --all and antigravity-cli enablement checks in skills/skillshare/references/plugins.md * Require known enabled state for native updates Require a matching EnabledKnown && Enabled inventory entry before updating copilot or antigravity-cli bindings, and align the localized plugin workflow pages with the English docs. - Skip copilot and antigravity-cli updates in Service.Preview and refuse reinstallation in applyNativeTarget when the binding is absent from native inventory or not known to be enabled - Cover absent native inventory in TestAntigravityCLIUpdateRespectsEnablement - Synchronize plugin update [name | --all] and Antigravity CLI update guidance across the ja, ko, zh-Hans, and zh-Hant copies of quick-reference.md and sharing-plugins.md |
||
|
|
efc492fd7a |
fix(update): audit tracked candidates before publication (#647)
* fix(update): audit tracked candidates before publication Prevent rejected tracked-update bytes from becoming readable through published target links by staging and auditing a private candidate before updating the active checkout. Publish the exact approved revision without a second upstream pull. Preserve sparse state, audit metadata, followed checkouts, and compatible edits. Recover staged and unstaged changes after failed forced publication, including deletions, renames, and sparse exclusions. Cover the boundary deterministically and align error messages and localized documentation. * fix(update): preserve local bytes across Git line-ending conversion Overlay dirty worktree bytes after staging candidate patches so core.autocrlf does not silently rewrite local edits. Cover staged, unstaged, and cancelled worktree changes while retaining their index state. Make permission-failure coverage explicitly POSIX-only and compare private checkout contents to their original bytes. Align red-team assertions with candidate rejection and register private temporary cleanup with the raw-write guard. * test(install): normalize checkout after configuring line endings Windows fixtures initially inherit core.autocrlf=true. Reset the isolated checkout after selecting the subtest setting so changing it to false does not turn untouched upstream files into local edits. |
||
|
|
5729eee088 |
feat(project): manage global registrations from the CLI (#603)
Add global project list, show, add, edit, sync and remove so setup scripts can register repositories without editing YAML or opening the dashboard. Keep dry runs read-only and report completed work when a target fails. Reuse the shared project service and sync runners. Include command docs, translations, completions and regression coverage. Quote overlap hints for their shell and log partial project syncs consistently with ordinary sync. |
||
|
|
2e22257520 |
fix(check): report failed update checks as failures (#643)
Return nonzero exit status for failed skill and agent checks while preserving JSON results. Propagate agent remote errors and prioritize dashboard check failures over the up-to-date summary. Add regression coverage and document exit status in all command reference locales. Related to runkids/skillshare-app#85. |
||
|
|
f3d23ccb0c |
feat(ui): draw skill trees as curved branches and animate folding
Nested packages sat in a grey inset panel and the folder tree drew one straight line per level, so neither showed which row a child hangs from. Both now use one faint spine under the parent with a rounded elbow into each child that stops short of its checkbox or icon, in an opaque colour so overlapping strokes leave no darker seam. Opening and closing a package or folder now grows and shrinks instead of jumping; collapsed content is inert while it shrinks, and reduced motion turns the animation off. Nested row actions line up with top-level rows, and a selected folder's block keeps straight sides. |
||
|
|
a96c5e55f2 |
fix(ui): warn that uninstalling a package removes its nested skills (#636)
* fix(ui): warn that uninstalling a package removes its nested skills Uninstalling a skill moves its whole folder to the trash, so the skills nested inside it go too, but the confirmation listed only the one skill. It now counts them in the title, marks the package row and adds a note. Selecting a package together with its nested skills (select all does this) also sent every name to the server, which trashed the package first and then failed on each child whose folder was already gone. The dialog now sends only the outermost selected skills. The success toast uses the same count as the title; it reported the server's per-name count, which read as fewer removals than confirmed. * fix(ui): keep nested skills of a linked folder root in uninstall A linked folder whose root is itself a skill can only be unlinked, so the server refuses to trash it. Collapsing the selection to that root dropped its selected children from the request and none of them were removed. A linked root no longer counts as holding its children: they are sent on their own and the root row does not claim to remove them. |
||
|
|
1a79971df4 |
feat(ui): show skill descriptions in the list and redesign package rows (#635)
* feat(ui): show skill descriptions in the list and let expand all open packages Reviewing a large library meant opening each detail page just to read what a skill does (#627, proposal in #631). The resources list response now carries each skill's and agent's frontmatter description, read in the same pass as manual-only, and the list shows it truncated under the name with the full text on hover. The read now decodes the frontmatter once for both fields. An unquoted ": " in a description is invalid YAML, so a failed decode falls back to the line scanner instead of dropping the description and manual-only. Editing a skill refreshes the list too, so the new description shows right away. Expand all / Collapse all also opens and closes nested-skill packages from #624; before, it only touched source and folder groups. Closes #627 * feat(ui): redesign list packages as an inset panel with a description switch The chevron-in-a-column nesting read as a second, unrelated table and gave no room for the descriptions #627 asks for. A count pill after the package name now opens its skills in one quiet inset panel, with deeper packages indented under a faint guide, so the hierarchy stays readable at any depth. The rest closes the gaps left against #627: a Show > Descriptions switch keeps the compact list one click away, keyboard focus reveals the full description, a search opens only the packages holding a match without overwriting what the user left open, and the selection bar says how many selected items are out of view. Only the name opens the detail page, since whole-row clicks were easy to trigger by accident once rows held more text. * fix(ui): keep linked-folder matches when searching a folder group Searching while grouped by folder took package ancestors only from skills whose folderOf matched the group key, but groupByFolder keys a linked folder's skills by the link name. Every skill in a linked folder fell out of its own group, so the header counted a match and drew no row. The ancestor pool is now each folder group's unfiltered skills, keyed by groupByFolder itself, which always contains the group's matches. * fix(ui): keep the .ss-r .nm.m rule the design token export reads Merging it into :is(.ss-r, .ss-nr) .nm.m removed the exact selector that scripts/design/tokens.py looks up for the row name type scale, so the design token tests failed in CI. |
||
|
|
a65414fa2f |
feat(ui): nest skill packages under their parent in the list view
A package that ships its own SKILL.md plus nested skills used to list every child as a peer, so one large package could fill the page and hide which skills belong together (#624, proposal in #625). The list view now puts each skill under the nearest skill whose folder holds it, within its current group. Packages start collapsed with a nested count, a search opens them so matches stay visible, and paging counts top-level rows so a package is never split across pages. Selection, select-all and bulk actions still work per skill. Closes #624 |
||
|
|
2d196c286b |
feat(check): show what changed in a skill since install (#623)
* feat(check): show what changed in a skill since install #622 made a hand-edited skill visible as modified and kept update from overwriting it, but there was still no way to see the edits before deciding to keep or discard them (#621). LocalChanges compares the skill's files with the hashes recorded at install, fetches the source at the recorded commit only when a recorded file changed, and diffs each file against it, so the diff shows the user's edits even after upstream moves on. A skill without a recorded commit is refused rather than diffed against HEAD, which would show upstream changes as the user's. It backs `check <name> --diff`, GET /api/resources/{name}/local-changes, and the dashboard: the Modified locally dialog on the skill page lists the diffs, and the Updates tab links a modified skill to it. update's skip hints now point at --diff. The route looks up the exact record and refuses non-local names, so a nested or ../ path cannot borrow a top-level skill's record or read outside the source. The file diff helper moved to textdiff so the CLI and server share it. * fix(install): diff local changes against the full installed commit entry.Version holds the short SHA from `rev-parse --short`. A targeted fetch needs the full SHA, so checkoutCommit fell back to fetching every branch and tag with --unshallow: viewing one edited skill could download the repository's whole history, and fail when the commit is only reachable by its full SHA. Use entry.Commit, the full SHA install records, when it is the same commit as Version; a Commit left from an older install, or a legacy record without one, falls back to Version. * test(install): keep checkout line endings when editing in LocalChanges test On Windows git checks the installed copy out with CRLF, so overwriting SKILL.md with LF content made every line differ and the expected diff failed. Edit the installed file in place, as an editor would, and compare the diff without carriage returns. |
||
|
|
7a7cceefab |
feat(update): flag and protect skills edited since install (#622)
* feat(update): flag and protect skills edited since install A skill installed from a repository and then edited by hand looked exactly like an untouched one: check said up to date, and update either reported it current or silently reinstalled it over the edit (#621). Install already records each file's hash, so check now compares a remote skill's files with those hashes and reports "modified", and update skips such a skill until --force (CLI single, --all/--group, dashboard). --force, the dashboard's Force Retry on the skipped row, or the confirmation on the skill page discards the edit and reinstalls. Two hash paths had to be made reliable for this to be trusted: the dashboard's folder-wide target edit refreshed hashes by basename, so a nested skill looked edited after skillshare's own change, and a skill whose files cannot be hashed now counts as modified instead of being overwritten. * fix(server): keep earlier local edits when a target change rewrites SKILL.md Changing a skill's targets in the dashboard rewrites its SKILL.md, and both the folder-wide and the single-skill handler then re-hashed the whole directory as the new baseline. An edit the user had already made to any file in the skill was absorbed into that baseline, so the skill stopped counting as modified and the next update overwrote the edit without --force. Both handlers now move only SKILL.md's recorded hash, and only when the file still matched it before the rewrite. skillshare's own target edit no longer reads as a local change, and an earlier edit still does. The whole-directory refresh helpers lost their last callers and are removed. |
||
|
|
330114bd6a |
fix(install): use frontmatter names for generic skill directories (#617)
* fix(install): use frontmatter names for generic skill directories Resolve meaningful defaults for generic containers while preserving explicit names, tracked paths, existing update destinations, and directory-based source selection. Fall back on unsafe or colliding names to prevent overwrites. Cover global and project installs, local and Git sources, naming overrides, tracked reports, collision safety, and updates. Closes #616. * fix(install): preserve names during source-based updates Resolve source-based updates against existing metadata before choosing destinations. Reject ambiguous or colliding update targets before writes, and reserve case-insensitive discovery names across filesystems. Add global/project regressions for local and Git source updates, legacy destinations, grouping, and upstream-added collisions. * fix(install): retain update ownership and selection checks Only allow a metadata-verified same-source update to bypass destination conflicts. Reject occupied fallback destinations for both local and Git sources unless overwrite is explicitly forced. Resolve destinations after skill selection so unrelated conflicts do not block filtered updates. Cover cross-source and unmanaged destinations, selected/excluded updates, and collision aborts in global and project modes. |
||
|
|
9dcd8d6e63 |
fix(ui): send project-target folder conflicts to the project editor
The Sync page offered "Stop syncing skills" for every target a folder conflict asked to stop. For a project target that button opened the skills-off dialog, whose preview and update calls the server refuses with 400 because project targets are configured through their project. Link those targets to their project's editor instead, and say so in the doctor docs. Fixes #612 |
||
|
|
dabe06a297 |
feat(targets): support OpenCode config directory accounts (#594)
* feat(targets): support OpenCode config directory accounts OpenCode v2 introduces support for custom configuration directories via OPENCODE_CONFIG_DIR, storing opencode.json and agent directories. Previously, skillshare rejected OpenCode accounts configured with a custom config directory because OpenCode was missing config_dir and was excluded from accountAgents in hooks and MCP services. - Define default config_dir (~/.config/opencode) in targets.yaml - Add OpenCode to account agents and register OPENCODE_CONFIG_DIR across hooks and MCP service handlers - Route OpenCode account MCP configuration to opencode.json inside the account config directory Resolves #593. * fix(targets): add OpenCode to schemas, completions and docs Add opencode to the targetConfig.agent enum and config_dir description in schemas/config.schema.json so configurations containing OpenCode account targets pass schema validation. Include opencode in the target --agent completion scripts for Bash, Zsh, Fish, PowerShell, and Nushell. Update website documentation and built-in skill references to describe OpenCode accounts and OPENCODE_CONFIG_DIR management alongside other account agents. * fix(targets): support plugin accounts and overlap Add OpenCode account recognition and OPENCODE_CONFIG_DIR resolution to the plugin service so plugin sync works for OpenCode account targets. Introduce TargetScanPaths to resolve cross-runtime scan paths using TargetConfig.Agent for account targets while retaining the account's own primary directory. Update path overlap detection, doctor checks, and LeakedSkills to warn about leaked skills for account targets. Update localized docs (zh-Hans, zh-Hant, ja, ko) and plugin references to document OpenCode account targets and OPENCODE_CONFIG_DIR. Update the available targets handler test to expect OpenCode's configDir. * fix(targets): isolate OpenCode accounts and scan paths Reject OPENCODE_CONFIG overrides when operating on OpenCode accounts and clear OPENCODE_CONFIG in the account execution environment so plugin operations target the account directory. In TargetScanPaths, remap config-root paths to the account directory while preserving external scan paths like Codex's ~/.agents/skills. * fix(plugins): use account target in OpenCode update probe Pass c.Target instead of a hardcoded "opencode" when probing whether OpenCode supports plugin updates, ensuring account-configured CLI binaries and account environment variables are used during preview. * test(mcp): isolate XDG_CONFIG_HOME in account test service Isolate XDG_CONFIG_HOME to the temporary test directory in agentAccountService to prevent runner environment variables from leaking into TestAccountEnvOverrideDoesNotPruneDefaultHome when testing OpenCode account shell overrides. --------- Co-authored-by: Willie <egg358358@gmail.com> |
||
|
|
7b952c02a9 |
fix(update): honor accepted findings in dashboard tracked updates (#599)
* fix(update): honor accepted findings in dashboard tracked updates The dashboard ignored findings accepted through the CLI, so a tracked repo that `update --force` let through was blocked again on the next dashboard update, and every dashboard Force Retry had to be repeated. Regular skill updates in the dashboard already honored and recorded acceptances, so only tracked repos differed. Tracked-repo updates now always honor accepted findings and record the ones an override lets through, for both callers. The opt-in option and the AuditGate flag it fed are removed, since no caller turns them off. Dashboard agent tracked repos passed an empty source dir, so a nested repo kept acceptances in its group folder under a different key than the CLI. They now pass the agents source, as the CLI does. * fix(update): reload the agents store after agent repo refresh Dashboard agent tracked repos now refresh metadata in the agents source, but a metadata change reloaded the skills store, leaving the agents store as it was until the next request. Reload the store that matches the source. Every caller of update.TrackedRepo now passes a source dir, so the empty-dir guard and its doc no longer describe a supported mode. The new agent test reuses the tracked fixture instead of rebuilding it. |
||
|
|
b69b3584d3 |
docs: clarify upstream skill and plugin tracking (#596)
* docs: clarify upstream skill and plugin tracking The built-in skill recommends --track --all, which install rejects. Show valid scoped workflows for following upstream collections and native plugins, distinguish update from sync, and explain selection flags before users hit their conflicts. * docs: clarify default-branch and retained-ref boundaries Qualify default-branch guidance for unpinned sources so update examples do not imply that an existing ref is cleared. Clarify mixed resource collections and filter modes alongside the workflow. * docs: address tracking workflow review feedback Remove rejected tracking/selection combinations from enterprise recipes in all five languages, document incompatible flags in the install tables, and state the selection errors explicitly. Keep the built-in skill concise by routing detailed upstream workflows to its references. |
||
|
|
829cbb176d |
fix(plugins): honor native OMP marketplace builds (#588)
* fix(plugins): honor native OMP marketplace builds Resolve OMP entries from their native marketplace when present and inspect package.json extension-only plugins. Keep target-specific source paths, marketplace identities and components so Claude adoption and hybrid manifests remain valid.\n\nShow separate builds in text discovery without changing logical JSON selection, and cover native precedence, invalid catalogs, manifestless and hybrid packages, global/project materialization and CLI output with regression tests. Update the plugin docs and existing translations.\n\nFixes #587 * fix(plugins): preserve native build metadata across consumers Keep OMP catalog versions in target discovery metadata and materialized marketplace entries so previews, cache guards, and native installs use the same version, including legacy catalog fallback. Cover catalog versions that differ from package manifests. Show each dashboard discovery target's own version, components, and source path instead of presenting a merged candidate as one build. * fix(plugins): avoid ambiguous shared build versions Derive candidate versions from agreement across supported targets and recompute after catalog overrides. No-Agent records and lists no longer inherit an arbitrary native build version; equal versions remain available and unknown versions stay unknown. Keep Cursor and Antigravity inventory target-specific when manifests disagree. Cover no-Agent preview and persistence, equal and unknown versions, catalog overrides, and mixed-manifest local inventories. Align English and translated plugin references. * fix(plugins): accept Pi extension manifests for OMP catalogs Reuse the existing OMP extension detector instead of requiring an OMP-specific manifest key. Native OMP accepts Pi extension manifests too, so the extra gate rejected supported native catalog entries. Cover Pi-only discovery with scoped package names and global/project materialization. Align the English plugin references and all existing translations. |
||
|
|
a2112ff681 |
feat(ui): keep all local or all remote in the pull conflict dialog
A pull with dozens of conflicting files meant clicking "Keep this version" once per file. Two buttons now set every file to the same side; individual files can still be switched afterwards. |
||
|
|
f6e7265414 |
feat(hooks): read hook code and scripts from files next to the config (#586)
* feat(hooks): read hook code and scripts from files next to the config
Code Agent extensions and command scripts lived as strings in
config.yaml, so a longer Pi extension or shell script was edited inside
a YAML block scalar with no editor support, and nothing type-checked or
linted it. A breaking change in an Agent's extension API went unnoticed
because the generated file is not part of any tsconfig.
A binding now takes `codeFile: path` instead of `code`, and a `files`
value may be `{from: path}` instead of the contents. Paths resolve
against the folder of the config that holds the entry and are read
every time the config is read, so the regular typecheck or shellcheck
can cover the real source. The revision and drift checks use the
resolved contents, so a file edited after a preview fails the apply as
stale.
Saving keeps the references: enable/disable, `hooks edit` and the
dashboard write `codeFile` and `{from}` back instead of inlining the
contents. JSON carries the contents plus `codeFile`/`filesFrom`, and an
entry echoed back with both must match the file, so an edit is refused
rather than silently dropped. Paths must stay inside the config folder
without symlinks, so a cloned project cannot pull in files from
elsewhere. References are taken from the YAML tree before one strict
decode, which keeps unknown-field checks and anchors shared across
bindings working.
Closes #585
* fix(hooks): refuse an emptied file sent with its reference
An empty files value stood for "contents not sent", so a file cleared in
the dashboard's Git YAML editor while its filesFrom reference stayed was
treated as unsent: the file was read again from disk and the save
reported success with the user's edit discarded.
A value read from {from: path} now keeps a dedicated mark until it is
resolved, so an empty value is compared with the file like any other
contents and refused when they differ. Code keeps using empty as unsent,
since a code binding can never be empty.
* fix(hooks): render referenced files as they are on disk
RenderNative rendered the contents the caller sent. The dashboard sends
the entry it loaded, so after a referenced file changed, View
configuration still showed the old contents while describing them as
what sync would write.
Render now drops sent contents for codeFile and {from} references and
reads every reference again, like preview and save. Draft, render and
config decoding are the only paths that take an entry, and all three now
resolve references.
* fix(hooks): accept file references in JSON entries and anchored values
Two inputs that config.yaml accepts failed to load elsewhere.
A `{from}` value carrying its own YAML anchor, aliased from another
binding, lost its anchor because the value node was replaced instead of
rewritten, so the alias no longer resolved. The node is now rewritten in
place with its anchor kept.
An entry JSON file (`hooks add|edit --file x.json`) could not write a
files value as `{"from": path}`, because JSON decoded straight into
string contents. Binding now decodes JSON files values as contents or a
reference, strictly, like YAML, so both entry formats and the Web API
take the same shapes.
|
||
|
|
70ceacca9b |
feat(mcp): serve an Executor skill index from mcp serve --http (#583)
* feat(mcp): serve an Executor skill index from mcp serve --http Executor apps load remote skills with wellKnownSkills, but could only reach Skillshare's skills through a copy in Git, so an edit saved with apply_skill_edit never reached them without a commit and push. With --http, mcp serve now also answers /.well-known/agent-skills/index.json and the files it lists, built from the same catalog and token, so Executor sees an edit on its next check. The format is the one Executor's loader reads, not the agent-skills discovery draft v0.2. One skill that loader rejects fails its whole load, so the index leaves out skills it would reject (non-text metadata, license or allowed-tools, files over 2 MB, non-UTF-8 files, file names decodeURIComponent cannot handle, names used twice) and lists them in --check and on stderr at startup. Versions are digests of each skill's paths and file digests, and the output is stable while nothing changes, because the loader reads the index twice and compares the bytes. An index over Executor's file or byte limits answers with an error naming --target. Manifest files now record their path within the listing skill: a parent's entry otherwise took its nested skill's paths, which would have listed SKILL.md twice. The index route repeats the MCP handler's DNS rebinding check for loopback servers without a token. Closes #574 * fix(mcp): accept proxied hosts with a token and use Executor's app accounts Behind the documented TLS proxy, a loopback mcp serve sees the proxy's public Host, and the DNS rebinding check answered 403 even with a valid bearer token. The index route had it, and so did the MCP endpoint through the go-sdk's localhost protection. A rebinding page cannot send the token, so both checks now apply only when no token is set; the docs say a tokenless server behind a proxy needs a loopback Host. The Executor example used the account form apps 0.0.1-beta.90 removed (`{ accounts }` and `ctx.accounts`, now typed as never). It declares the token as an app account and reads ctx.app.accounts, in all five locales. The wellKnownSkills loader itself is unchanged at the v2 head (b98424d9). |
||
|
|
8f55e58999 |
feat(ui): edit other Agents' MCP options in the server dialog (#580)
* feat(mcp): write per-Agent native fields from the source Some MCP settings have no portable form, such as Codex's default_tools_approval_mode or Claude Code's per-server timeout. Sync kept them only while an entry survived, so a recreated entry (for example after a rename) lost them, and every machine needed them added by hand again. A server's `options` now holds these fields per Agent, and sync writes them as given. The field ledger that piOptions used is extended to every Agent, so sync removes an option it wrote once the source drops it, and reports a conflict when the Agent changed it since. Options are refused when they could break the entry or leak a secret: fields Skillshare writes (including portable names some Agents rename), values a native format cannot write back (null, arrays of tables), options.pi (piOptions covers Pi), disabled entries, and credentials as plain text, since the Agent's file may be committed. The dashboard has no editor for options yet but keeps them on save. Refs: #578 * fix(mcp): close gaps in per-Agent options found in review - A credential option passed if it merely contained "${", so text around a reference, or another Agent's syntax, was written out as plain text. It must now be a whole-value reference in the Agent's own syntax, and Codex and Antigravity options, which read no references, cannot hold one. - The Kilo Code project check looked only at portable env references, so an {env:VAR} in options reached a project file Kilo refuses to load. - A replace resolution left the dropped-option conflict in place, so the plan stayed blocked; replace now lets the source win and prunes those fields. - Adopting an entry that already matched skipped writing its options until a later sync; the takeover now writes them. * fix(mcp): refuse numeric credentials and constrain option values in the schema - YAML reads an unquoted secret such as clientSecret: 123456 as a number, which the credential check skipped, so sync wrote it out as plain text. A mapping under a sensitive name such as credentials also hid its plainly named fields. Sensitivity now carries down through mappings and lists, and a number under a sensitive name is refused. - The schema allowed any value inside an Agent's options, so editors accepted null or nested lists that LoadSource then rejects. A recursive optionValue definition now matches the runtime rules. * feat(ui): edit other Agents' MCP options in the server dialog The dialog only carried options through unchanged, so setting a field such as Codex default_tools_approval_mode or Claude Code timeout still meant editing mcp.yaml by hand. One Agent picker with a JSON editor keeps the dialog short however many Agents are ticked; ticked Agents come first, and an unticked Agent that still has settings stays listed so they can be cleared. Anything but a JSON object blocks saving; the server checks the fields themselves. * fix(ui): offer every Agent in the options picker and document the editor - The picker listed targets from the order, which includes accounts, so a ticked account produced an options key the server refuses, and the Agent the account writes to was not offered. Options are keyed by Agent, so every Agent but Pi is offered, ticked ones first. - An unticked Agent left the list as soon as its text was emptied, so its settings could not be cleared and retyped. - The editor was hidden on the Skillshare tab although its server takes options too, and hidden invalid text could block Save there. - The docs still said the dashboard had no options editor. * fix(ui): open the options editor on a ticked account's Agent The picker ranked Agents by the ticked targets, which hold account names, so a server reaching only codex-work opened the Claude editor and what was typed there was saved under options.claude. The MCP page now provides each account's Agent, and the ranking uses it. The retype test now asserts the saved options rather than the editor staying visible. * fix(ui): show the server's refusals of Agent options in the dashboard language The dialog leaves field rules to the server, but its refusals (a field Skillshare writes, a plain-text credential, a credential in an Agent that reads no references, an unsupported value) reached the dialog in English. describeError now maps each to a translated message naming the Agent and field. The other option errors cannot come from the dialog: it offers no Pi, no non-Agent key, and no options on an off switch. |
||
|
|
578790c416 |
feat(mcp): write per-Agent native fields from the source (#579)
* feat(mcp): write per-Agent native fields from the source Some MCP settings have no portable form, such as Codex's default_tools_approval_mode or Claude Code's per-server timeout. Sync kept them only while an entry survived, so a recreated entry (for example after a rename) lost them, and every machine needed them added by hand again. A server's `options` now holds these fields per Agent, and sync writes them as given. The field ledger that piOptions used is extended to every Agent, so sync removes an option it wrote once the source drops it, and reports a conflict when the Agent changed it since. Options are refused when they could break the entry or leak a secret: fields Skillshare writes (including portable names some Agents rename), values a native format cannot write back (null, arrays of tables), options.pi (piOptions covers Pi), disabled entries, and credentials as plain text, since the Agent's file may be committed. The dashboard has no editor for options yet but keeps them on save. Refs: #578 * fix(mcp): close gaps in per-Agent options found in review - A credential option passed if it merely contained "${", so text around a reference, or another Agent's syntax, was written out as plain text. It must now be a whole-value reference in the Agent's own syntax, and Codex and Antigravity options, which read no references, cannot hold one. - The Kilo Code project check looked only at portable env references, so an {env:VAR} in options reached a project file Kilo refuses to load. - A replace resolution left the dropped-option conflict in place, so the plan stayed blocked; replace now lets the source win and prunes those fields. - Adopting an entry that already matched skipped writing its options until a later sync; the takeover now writes them. * fix(mcp): refuse numeric credentials and constrain option values in the schema - YAML reads an unquoted secret such as clientSecret: 123456 as a number, which the credential check skipped, so sync wrote it out as plain text. A mapping under a sensitive name such as credentials also hid its plainly named fields. Sensitivity now carries down through mappings and lists, and a number under a sensitive name is refused. - The schema allowed any value inside an Agent's options, so editors accepted null or nested lists that LoadSource then rejects. A recursive optionValue definition now matches the runtime rules. * docs(mcp): name every Agent whose options cannot hold a credential The docs named only Codex and Antigravity, but Claude Desktop, Goose, Junie, LM Studio and Warp read no environment references there either, so a source such as options.goose.apiKey: "${GOOSE_KEY}" looked valid while LoadSource refused it. |
||
|
|
4a5dacac6b |
feat(mcp): let mcp serve clients edit listed skills (#577)
* feat(mcp): let mcp serve clients edit listed skills A remote assistant that reaches skills only through `mcp serve` could read them but had to hand every change back to a person (#540). Skills listed in `mcp.serve.edit.skills` can now be changed one existing file at a time: `preview_skill_edit` writes nothing and returns the full diff, the audit result and a preview id, and `apply_skill_edit` saves exactly that preview, once. Every save is tied to the revision the client read, so a file changed in the meantime fails with a conflict instead of being overwritten. Skills that `update` would overwrite stay read-only: tracked repos, upstream installs, and skills nested inside an install. The skill is validated and audited in a temporary copy before anything is written; as with install, a finding at or above `audit.block_threshold` refuses the edit unless it was accepted with `--force`, and findings are marked new or existing so the client can see what its change adds. The write is atomic, keeps the file mode, goes through sourcefs link checks, and is recorded in the operation log. Native Skills extension clients still see the edit tools, since the extension cannot write. * fix(mcp): close gaps in mcp serve skill editing Codex review of #577 found five ways the edit tools could do more, or record less, than they promise: - A blank pattern in mcp.serve.edit.skills was dropped by the filter, and no patterns match every skill, so `skills: [' ']` made every skill editable. Blank patterns are now a config error. - preview_ttl was only checked when skills were listed, so the dashboard could save an invalid value that later stopped mcp serve. It is now checked whenever edit is present. - Apply compared only the edited file's digest, so two previews of different files, or an edit made elsewhere in between, could save a skill state that was never validated or audited. Apply now checks and audits the skill again as it will be saved. - Edit audits were not written to audit.log, so blocked attempts left no trace in `skillshare log --audit`. Each preview and apply audit is now logged like install's audit gate. - The writability check migrated legacy metadata, so a "writes nothing" preview (and --check) could rewrite .metadata.json and delete registry.yaml. It now reads the store without migrating; a legacy registry.yaml keeps every skill read-only, and a legacy sidecar marks its skill as installed. |
||
|
|
fdca87a1ff |
feat(mcp): mark the mcp serve skill tools read-only
Clients and gateways that gate tools by annotation treat a tool without readOnlyHint as one that changes data. Executor v2 does, so every list_skills or read_skill call could stop for approval even though both only read the skills source. Marking them read-only lets those clients call them without asking. Refs: #540. |
||
|
|
4e3076df3f |
docs: describe token helper auth and the update fetch warning
0.25.12 changes two things users can notice that the docs never covered. Token auth now goes through a host-scoped credential helper (#566), so the token stays out of ps and git skips the user's own helper or keychain for that host. Grouped updates now warn when a repository cannot be fetched (#570), with credentials removed from the URL and error. Note both on the install and update pages in every locale. |
||
|
|
0c69dad6c2 |
fix(update): stop trusting tree hashes for skills with symlinks (#573)
* fix(update): stop trusting tree hashes for skills with symlinks Install copies a file symlink's target, but git stores the link as a blob holding the target path, so a skill's subtree hash stays the same when a target outside the folder changes. Batch update skipped such skills as current and check reported them up to date, leaving stale content installed. A subdir that holds a symlink now has no usable tree hash: install records none, update falls back to a reinstall, and check reports the skill as update_available. Check lists entries with -z so links with non-ASCII names are not missed behind git's path quoting. Closes #572 * fix(check): split ls-tree -z output on NUL only parseLsTreeOutput also split on newlines so the old newline fixtures kept passing, but with -z a newline belongs to the path. A directory or symlink whose name holds a newline was cut in two, and a symlink under it was missed. The fixtures now use NUL-terminated entries like real -z output. Also writes both shortcut comments in the repo's one-line form and uses testutil.RunGit in the new update test instead of a raw exec call. * docs(update): note that skills with symlinks always reinstall The update page says grouped updates leave skills whose folder did not change alone. A skill with a symlink no longer has a usable tree hash, so it is reinstalled whenever the upstream commit moves. State that exception on the English page and its four translations. |
||
|
|
c4b183f6ef |
fix(update): read store in batch fast path (#569)
* fix(update): read store in batch fast path The grouped update fast path read per-skill .skillshare-meta.json sidecars. The migration to .metadata.json deletes those files, so the check never matched and update --all reinstalled every skill on every run. Read the central store instead, and refresh version and commit there instead of writing a sidecar. Require SKILL.md on disk so the update restores a deleted install. --force now skips the check and reinstalls everything. * test(update): cover legacy grouped metadata keys Grouped skills can still be stored under a basename key with a Group field. Check that the current-skill lookup finds that entry and that the version refresh updates it in place instead of adding a new key. * feat(update): report up-to-date skills The summary lumped skills the fast path skipped in with failures. Count them as "up to date" in the summary and as up_to_date in --json output. * feat(update): log up-to-date count in oplog The operations log recorded updated, skipped, pruned, and blocked counts for update but not the new up-to-date count. Record it as up_to_date so the log matches the summary and --json output. * test(update): cover up-to-date batch skills update --all reinstalled every grouped skill on every run. Check that unchanged skills report as up to date in global and project mode, that an upstream change updates only the changed skill, and that --force reinstalls everything. * docs(update): describe up-to-date batch skills Describe how update --all and --group leave unchanged skills alone and report them as up to date, the up_to_date JSON field, and the new meaning of --force. * fix(update): keep lock pins on tree-only matches A tree-hash match after upstream HEAD moved refreshed the full commit in the store. A project lock pin follows that field, so an unrelated upstream commit rewrote skills.lock.json even though the skill's files did not change. Refresh only the version, as the old sidecar refresh did. check compares the version, so it still reports the skill as up to date. The help text, docs and skill reference said --all and --group skip current skills. The skip lives in the batch path, which runs for any update of several skills, and a single skill always takes the reinstall path. Say that instead. * fix(update): load the store once per batch Each skill skipped as current reloaded .metadata.json before comparing its version, so an unchanged batch reparsed the whole store once per skill. Compare the version already read from the snapshot first, and refresh the skills that do lag HEAD with one load and save after the loop. An unchanged batch of 400 skills drops from about 520ms to 64ms. The help text and docs said any update of several skills takes this path. In global mode, named and glob targets carry no metadata and reinstall on the standalone path. Name what does take it: --all, --group, and several skills named in project mode. --------- Co-authored-by: Willie <egg358358@gmail.com> |
||
|
|
fe92a50b1b |
fix(upgrade): preserve UI assets for the pre-upgrade version (#571)
An already-running UI server still reads lazy-loaded chunks from its versioned cache after the CLI binary is upgraded. Preserve the pre-upgrade version alongside the downloaded version so upgrade cleanup does not break page navigation. Add a multi-version cleanup regression test and document the retention policy and accepted consecutive-upgrade limitation across all website locales. |
||
|
|
6cc790965a |
feat(mcp): pin tool definitions in mcp check --live and warn when they change (#560)
* feat(mcp): pin tool definitions in mcp check --live and warn when they change An Agent reads a server's tool names and descriptions as instructions, and a server can change them after the user reviewed it (a "rug pull"). The MCP spec 2026-07-28 local-server security guide notes the protocol does nothing about this and advises clients that re-prompt on a changed definition; most clients do not. `mcp check --live` now records one SHA-256 per tool over name, description and inputSchema the first time a server answers, in mcp/tool-pins.json of the state directory, keyed by scope and server. Later probes report a changed or new tool as a `tool-change` warning and a removed one as info, until `--accept-tools` (only with --live) records the current definitions. Without --live the pins are never read or written. An unreadable pin file is never rewritten, since it may hold other servers' pins. The dashboard words both warnings in every locale. * fix(mcp): keep schema numbers exact, name the scope to accept in, and log acceptance Codex review on #560: - Tool definitions were decoded through float64 before hashing, so a schema bound changed from 9007199254740992 to 9007199254740993 hashed the same and raised no warning. Numbers now keep their literal text. - The suggested accept command left out -g/-p; run from inside a project after a global check, it would accept the project's server instead. It now names the scope of the check that warned. - Accepting changed tool definitions approves a possible rug pull, but left no record. --accept-tools now writes an operations log entry with the servers it accepted. * fix(mcp): log a tool acceptance only when its baseline was saved Codex review on #560: when comparing succeeded but writing tool-pins.json failed (lock held, state directory unwritable), the report kept its "accepted" finding, so the operations log listed the server as accepted with status ok although its old pins remained. Such a server is now logged under `failed` and the entry's status is error. * fix(mcp): log first pins as accepted and report errors in the accept entry Codex review on #560: - A check that accepted one server but found an error on another still logged status ok, while the command itself exited 1. A server with an error is now listed under `failed` and the entry's status is error. - --accept-tools on a server's first probe records its baseline, but the finding says "pinned", so the entry listed nothing as accepted. A first pin counts as accepted now. * chore(release): request version 0.25.11 The MCP pin checks (#559, #560) and the backup staging fix (#558) ship as 0.25.11 instead of the computed 0.26.0, per maintainer decision. Release-As: 0.25.11 |
||
|
|
4382a8c8c8 |
feat(mcp): warn when a local server runs an unpinned package or image (#559)
* feat(mcp): warn when a local server runs an unpinned package or image An entry such as `npx -y some-mcp` fetches whatever the registry serves at every Agent start, so a compromised release reaches the machine on the next launch without anyone noticing. The MCP local-server security guide (spec 2026-07-28) names version pinning as a baseline and points out that a container tag can move while a digest cannot. `mcp check` now adds a `pin` finding from command and args alone, with no network: npx, uvx and `pipx run` with no version or `latest`, and `docker`/`podman run` with no tag or `latest`, are warnings; an image pinned by tag but not by `@sha256:` digest is info. Value-taking run flags are skipped so hardened lines such as `--cap-drop ALL` are not read as the image. Warnings keep the exit code at 0, as before. * fix(mcp): check every npx package, word pin warnings in the dashboard, document them in doctor Codex review on #559: - npx accepts --package more than once and fetches each one, but the pin check stopped at the first, so `npx --package=lodash@4.17.21 --package=cowsay cowsay` passed although cowsay is unpinned. Every --package/-p value is now checked. - The dashboard words env, command and dns findings in the viewer's language but showed the new pin warning in English. It now has its own sentence for an unversioned package and for an untagged image in every locale. The dashboard hides info findings, so the digest note needs none. - doctor promotes check warnings, so an unpinned launcher now shows there; its docs list the pin check. The mcp check runbook gains a step for the pin check. * fix(mcp): check uvx --with packages and ignore environment markers in pins Codex review on #559: - uvx --with (-w) installs each named package beside the tool, so an unpinned one moves just like an unpinned tool, but the pin check skipped it as a flag value. Every --with requirement is now checked next to the --from or command package. - In `--from 'mcp-server; python_version == "3.12"'` the == belongs to the environment marker, not to the package, yet it read as a pinned version. Only the requirement before ';' is inspected now. * fix(mcp): treat only exact versions as pins, and pin the docs' own examples Codex review on #559: - npm resolves a dist-tag (`@next`) or a range (`@^1.2.3`) to whatever release matches at launch, yet any selector but `latest` passed as pinned. An npx package is pinned now only by an exact semver version, the rule Pi bindings already use. - `==1.*` is a prefix match in Python version specifiers, so `uvx --from 'mcp-server==1.*'` can pick up a new 1.x release; it is no longer read as a pin. - The Context7, Playwright and filesystem examples in the docs and the built-in skill launched unpinned packages, so copying them produced the new warning at once. They now name the current registry version. |
||
|
|
cf5783ee88 |
fix(backup): reclaim abandoned staging directories safely (#558)
* fix(backup): reclaim abandoned staging directories safely Interrupted backup copies leave hidden .snapshot-* directories that retention cleanup never discovers. Hold a shared process lock during copies and use an exclusive cleanup lock to reclaim orphaned staging without deleting active work. Share retention planning with non-mutating dry runs, report busy manual cleanup in the CLI and API, and align dashboard messages and translated docs. Cover process interruption and automatic cleanup with regression tests and an executable E2E runbook. Closes #556 * fix(backup): address staging cleanup review findings Keep cleanup reachable for staging-only dashboard backups and disclose abandoned incomplete directory deletion in every confirmation locale. Cover the empty-state confirmation with a regression test. Read staging entries before a read-only lock probe so first-use dry runs cannot count a newly active writer. Refresh entries only under the acquired lock and classify both backup-root reads in the filesystem guard. * fix(backup): align cleanup guides and doctor status Disclose abandoned incomplete directory removal in backup help, completion descriptions, and the daily backup guide with all four translations. Add help and completion regressions so these consumers stay aligned with cleanup. Keep the doctor Backups row visible when only the persistent lock remains after cleanup, preserving the existing JSON check and covering the lock-only state. * fix(backup): complete cleanup documentation disclosure Cleanup reclaims abandoned incomplete directories regardless of retention, so retention-only summaries understated the action. Align the backup synopsis, option table and previews, related workflow guides, built-in reference and all four website translations. Scope the retention defaults to completed snapshots. |