@xevy/skills 0.1.68 → 0.1.70

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -2,7 +2,7 @@
2
2
  "name": "xevy",
3
3
  "displayName": "Xevy",
4
4
  "description": "Native Xevy SDLC and Features workflows through MCP: plan, implement, review, record QA evidence and publish verified screenshot guides.",
5
- "version": "0.1.68",
5
+ "version": "0.1.70",
6
6
  "author": {
7
7
  "name": "Xevy"
8
8
  },
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@xevy/skills",
3
- "version": "0.1.68",
3
+ "version": "0.1.70",
4
4
  "description": "Install Xevy agent skills and secure per-agent MCP transports for Cursor, Claude Code, Codex, and other coding agents.",
5
5
  "type": "module",
6
6
  "keywords": [
@@ -1,26 +1,42 @@
1
1
  ---
2
2
  name: deploy
3
- description: Deploy a per-task preview or the project's main environment via Xevy (Coolify). Usage:/xevy:deploy [preview|main] [task]. Bare invocation targets the in-context task's preview, else main.
4
- argument-hint: "[preview|main] [task]"
3
+ description: Deploy a task preview, the shared staging/development environment, or an explicitly registered production application through Xevy. Confirm production by its saved destination and report the deployment's actual outcome.
4
+ argument-hint: "[preview|main|production] [task or repository]"
5
5
  disable-model-invocation: true
6
- allowed-tools: "Read Bash mcp__xevy__deploy_preview mcp__xevy__deploy_main mcp__xevy__get_deploy_status mcp__xevy__find_task"
6
+ allowed-tools: "Read Bash mcp__xevy__deploy_preview mcp__xevy__deploy_main mcp__xevy__get_deploy_status mcp__xevy__find_task mcp__xevy__list_repositories mcp__xevy__project_deployment_target mcp__xevy__project_deployment_redeploy mcp__xevy__project_deployments_list mcp__xevy__project_deployment_logs"
7
7
  ---
8
8
 
9
- # /xevy:deploy — deploy a preview or main
9
+ # /xevy:deploy — deploy a preview, staging or production application
10
10
 
11
- Trigger a Coolify deploy through the Xevy MCP. `$ARGUMENTS` may start with `preview` or `main`; the rest is a task hint.
11
+ Resolve the requested environment and repository, confirm the consequential action, then follow the build to its recorded outcome. Deployment authorization and PR-merge authorization are separate actions.
12
12
 
13
13
  ## Resolve target
14
- - First word `preview` → per-task preview. `main` → project main. If unspecified (smart default): if a task is in context (`.xevy` session, an open task, or a task hint) → **preview**; otherwise → **main**. If genuinely ambiguous, ask the user.
14
+ - `preview` selects a per-task preview. `main` selects shared staging/development. `production` selects the repository's explicitly registered production application. With no environment supplied, use a task preview when a task is in context and shared staging/development otherwise. Preserve an explicit production request throughout the workflow.
15
+ - Resolve repository identity from the session, linked task or `list_repositories`. Ask when multiple repositories remain plausible. Branch names, hostnames and provider UUIDs from chat do not establish a production target.
16
+ - In a Dev Session, activate `repositories` capabilities to discover `project_deployment_target`, `project_deployment_redeploy`, `project_deployments_list` and `project_deployment_logs`. These use Xevy's authenticated service path; provider credentials stay outside the agent shell.
15
17
 
16
18
  ## Preview
17
19
  - Resolve the task (`mcp__xevy__find_task` by code). A pushed feature branch must exist. Call `mcp__xevy__deploy_preview({ taskId })` using recorded evidence, or optionally supply its exact pushed `branch` when not yet recorded. Conflicting branch input is rejected; never rewrite repository identity to force deployment.
18
20
  - Association is automatic: branch-only legacy tasks check linked deployment-enabled repositories; existing scoped records and exclusions remain authoritative. No repository selector is required. Every selected branch must be verified exactly; absent branches, inaccessible repositories, missing credentials or unavailable verification are skipped. Never substitute a default branch.
19
21
  - Report actual `queued`, `partial` or `skipped` status, URLs and each repository's reason. Partial results mean only independently verified repositories queued. If nothing queued, do not claim a build started; explain the recorded reasons.
20
22
 
21
- ## Main
22
- - Call `mcp__xevy__deploy_main` for the current project. It deploys the project's REMOTE default branch and never pushes git branches; warn that unmerged local work won't appear until landed via PR. Report the main URL.
23
+ ## Main: shared staging/development
24
+ - Obtain approval for shared staging/development, then call `mcp__xevy__deploy_main` for the current project. It deploys the project's REMOTE default branch. Local changes become available after their PR lands remotely. Report the main URL and follow it with `get_deploy_status`. Production requests use the production workflow below.
25
+
26
+ ## Production
27
+
28
+ 1. Confirm the requested repository and production release are authorized. If merging is needed, complete the separately approved PR merge and verify its remote receipt first.
29
+ 2. Call `project_deployment_target` with the selected project, `repositoryId` and `environment: "production"`. Read the saved application name, `canRedeploy` and opaque `confirmTarget`.
30
+ 3. Obtain approval for this exact production destination. The primary Dev Session shows the production application and target confirmation. Nested agents return release actions to the primary session. An earlier blanket approval for other work does not approve this destination.
31
+ 4. Call `project_deployment_redeploy` with the same scope, `environment: "production"`, the exact application name as `confirmName`, and the returned `confirmTarget`. A changed destination requires a fresh lookup and approval.
32
+ 5. Follow the returned deployment with `project_deployments_list`; use `project_deployment_logs` for bounded diagnostics. A queued response confirms dispatch. Report completion when that deployment reaches `succeeded` and any available application health check passes. Report `failed` and `cancelled` outcomes accurately.
33
+
34
+ - `target_missing` requires restoring the verified production link to the existing application before deployment. Report the missing registration and continue independent authorized work.
35
+ - `destination_read_only` means the registered connector supports reads. `hosting_forbidden` means the saved provider scope cannot manage the resource. Report the specific access requirement and preserve the existing scope boundaries.
36
+ - An unavailable deployment receipt leaves dispatch uncertain. Inspect recent deployments before retrying. `deployment_busy` requires following the active build before another request.
37
+ - Production deployment uses the existing registered application. Provisioning replacements, changing configuration or ENV, restarting supporting services and bypassing ownership checks require their own authorized workflows.
23
38
 
24
39
  ## Report progress
25
- - Queued deploys are async (queued → building → live). Use `mcp__xevy__get_deploy_status` to report current state. Tell the user the URL may take a few minutes to become reachable.
40
+ - Preview and shared staging/development deploys use `get_deploy_status`. Production uses `project_deployments_list` and `project_deployment_logs`. A queued or running build remains open until its terminal outcome or a stated observation/access limit.
41
+ - Give the repository, environment, deployment ID and actual state. Include the URL when the service returns one. Separate successful dispatch from application verification.
26
42
  - If the server reports previews are disabled, say so and stop. Never push protected branches (master/main/develop/staging) to make a deploy reflect local changes.
@@ -39,6 +39,8 @@ Call `mcp__xevy__list_task_decisions` before implementing. Existing unanswered P
39
39
 
40
40
  ### Worktree protocol
41
41
 
42
+ - In a task Dev Session, reuse the task branch and worktree already registered at startup. Register an additional task through `register_dev_session_branch` only when needed, and edit its returned directory after verified success. Retry a failed registration using the same `branchId`; listing branches or waiting does not complete registration.
43
+
42
44
  - Record the current branch and directory. Derive a canonical branch `xevy/<TASK-ID>-<slug>` using the task code or the first eight characters of its UUID. Reuse an existing task branch; never create a second branch for the same task.
43
45
  - Create or attach a git worktree in a sibling directory and change into it. When resuming, first merge the latest source branch (for example `git merge origin/main`). Do all work in the task worktree.
44
46
  - Verify the absolute worktree path with `git rev-parse --show-toplevel` and its branch with `git branch --show-current` before editing. Keep the task code, UUID, title, absolute worktree path, and branch in the session context. On resume or after changing directories, verify and refresh that context before continuing.
@@ -3,7 +3,7 @@ name: merge-pull-request
3
3
  description: Evaluate a GitHub pull request, repair merge conflicts and PR-caused failures on its head branch, verify it, and merge it when ready. Use whenever the user asks to review and merge a PR, resolve a PR conflict before merging, or replace the former merge --cloud workflow. Raise remaining issues and blockers with evidence, then ask whether to create Xevy tasks or explicitly disregard them before proceeding.
4
4
  argument-hint: "[pull-request URL|number|branch]"
5
5
  disable-model-invocation: true
6
- allowed-tools: "Read Grep Edit Bash mcp__xevy__find_task mcp__xevy__list_tasks mcp__xevy__get_project_workbreakdown mcp__xevy__create_task"
6
+ allowed-tools: "Read Grep Edit Bash mcp__xevy__find_task mcp__xevy__list_tasks mcp__xevy__get_project_workbreakdown mcp__xevy__create_task mcp__xevy__read_dev_session_pull_request mcp__xevy__prepare_dev_session_pull_request_review mcp__xevy__merge_dev_session_pull_request mcp__xevy__list_dev_session_branches mcp__xevy__push_dev_session_branch"
7
7
  ---
8
8
 
9
9
  # /xevy:merge-pull-request — evaluate, repair, and merge a pull request
@@ -13,14 +13,16 @@ Review the exact pull request, repair merge conflicts and PR-caused failures on
13
13
  ## 1. Resolve the pull request
14
14
 
15
15
  - Resolve `$ARGUMENTS` as a GitHub pull-request URL, pull-request number, or head branch. When it is empty, use the pull request associated with the current branch only when exactly one open pull request matches.
16
- - Verify GitHub CLI authentication and repository identity before making changes.
17
- - Read the pull request with `gh pr view`, including its number, URL, state, draft state, head and base branches, mergeability, merge-state status, review decision, checks, commits, and changed files.
16
+ - In a Dev Session, activate the `dev_session` capability group and use `read_dev_session_pull_request` with the linked `repositoryId` and PR number. Read metadata, files, commits, reviews and checks; follow every `nextPage` and treat `truncated` as incomplete evidence. The session's pinned GitHub connection stays inside the toolbox; shell GitHub authentication is unnecessary.
17
+ - Save `headSha`, the current `baseSha`, and the base branch from fresh metadata. `data.base.sha` may describe an older recorded base; use the top-level `baseSha` for the current target tip.
18
+ - On a workstation, verify GitHub CLI authentication and repository identity before making changes. Read the pull request with `gh pr view`, including its number, URL, state, draft state, head and base branches, mergeability, merge-state status, review decision, checks, commits, and changed files.
18
19
  - Stop with a clear blocker when the pull request is closed, already merged, missing, or belongs to a different repository than the checkout.
19
20
  - Record the exact head repository and branch. Repairs belong on the pull-request head branch; protected base branches receive changes through the pull request.
20
21
 
21
22
  ## 2. Evaluate the change
22
23
 
23
24
  - Fetch the base and head refs, inspect the complete pull-request diff, and read the relevant callers, tests, migrations, configuration, and established patterns.
25
+ - In a Dev Session, use `prepare_dev_session_pull_request_review` to obtain a detached checkout of the exact PR head and current base. Run verification in its returned directory. Reconcile the returned SHAs with the reviewed evidence before continuing.
24
26
  - Check the pull request against its stated goal and repository instructions. Focus on correctness, security, data integrity, backward compatibility, operational safety, and test coverage.
25
27
  - Run the project-relevant type-check, lint, test, and build checks. Separate failures caused by the pull request from known baseline failures when evidence permits.
26
28
  - Inspect required GitHub checks, requested reviews, branch protection, draft state, and merge queue requirements.
@@ -36,6 +38,7 @@ Review the exact pull request, repair merge conflicts and PR-caused failures on
36
38
  - Repair verification failures introduced by the pull request or by the conflict resolution when the fix is clear and stays within the pull request's scope.
37
39
  - Pause for a product decision when a conflict has multiple valid outcomes with materially different behavior.
38
40
  - Stage and commit the repairs with an evidence-based message, then push the pull-request head branch. Never push a protected base branch directly.
41
+ - In a Dev Session, repair in a registered worktree matching the PR's head branch and publish through `push_dev_session_branch`. A detached review checkout is for verification. If the head branch cannot be published through the session's registered branch authority, report that limitation before attempting a repair push.
39
42
  - Rerun affected local checks and refresh the GitHub pull-request state after the push. Wait for required checks through GitHub's supported check-watching command when they are running.
40
43
 
41
44
  ## 4. Handle remaining findings
@@ -60,8 +63,11 @@ Review the exact pull request, repair merge conflicts and PR-caused failures on
60
63
 
61
64
  - Refresh the pull-request state immediately before merging. A ready pull request is open, non-draft, conflict-free, free of unresolved blockers, and compliant with required checks and repository protection.
62
65
  - Follow the repository's established merge method. Infer it from repository settings and recent merged pull requests; use squash when the repository permits multiple methods and no convention is evident.
63
- - Merge with `gh pr merge` and keep branch deletion separate from this command.
66
+ - In a Dev Session, call `merge_dev_session_pull_request` with the session repository, PR number, `expectedHeadSha`, `expectedBaseBranch`, `expectedBaseSha`, and selected `mergeMethod`. Use the exact reviewed head and current base evidence. The primary session requests approval for this specific merge; nested agents return the action to the primary session.
67
+ - If the head, base branch or current base tip changed, refresh the PR and review the affected changes before obtaining fresh approval. Pending or failed checks, unresolved mergeability and GitHub protections remain release gates. Credentials stay inside the toolbox.
68
+ - Merge with `gh pr merge` on a workstation and keep branch deletion separate from this command. Use the reviewed head SHA with the CLI's matching-head option when available.
64
69
  - Confirm the remote pull request reports a merged timestamp and merge commit. Report the pull-request URL, merge method, commit, repaired conflicts, verification results, and any approved disregards or follow-up issues.
70
+ - An unavailable merge receipt leaves the outcome uncertain. Read the PR before retrying; a timed-out request may have completed. An already-merged receipt confirms existing state without another merge request.
65
71
 
66
72
  ## Boundaries
67
73
 
@@ -46,6 +46,20 @@ Review the exact PR and report evidence-based findings. Review remains **read-on
46
46
  - Run relevant tests, lint, type-check, and build checks only where safe in the permitted working directory; avoid commands that rewrite tracked files or require secrets. Select checks from the risk-based matrix, prioritize high-impact paths, and run independent checks concurrently when safe and supported. If not run, explain why. Separate PR-introduced failures from baseline/infrastructure failures where evidence supports it. Never claim there are no code issues solely because CI is green.
47
47
  - Immediately before the verdict, re-read the PR's head and base SHAs. If either changed during review, refresh affected diff, mergeability, checks, and verification or mark the verdict stale/inconclusive; never present an earlier SHA's results as current.
48
48
 
49
+ ### Migration-history safety (when triggered)
50
+
51
+ - **Trigger on behavior, not filenames.** Apply this check when the PR changes migrations, schema, journals or equivalent history metadata, migration configuration, runner code, or startup/deployment migration execution. Include relevant unchanged runner and guard code in the review; a schema or deployment change can alter upgrade safety without adding a migration file.
52
+ - **Establish the actual runner contract from evidence.** Trace the configured entrypoint, resolved runner version/implementation, and startup/deployment wrappers. Identify the authoritative repository journal or equivalent, the database applied-history table/ledger, and the exact eligibility and ordering rules. Distinguish individual migration identity tracking from a latest-timestamp/version watermark; determine strict versus inclusive comparisons, tie handling, checksums/content hashes (including whether they are enforced), dependencies, and reconciliation/startup guards. Cite the relevant code/configuration and source/current-target SHAs; do not impose one framework's rules on another. Filenames and numbering alone never establish eligibility or a defect.
53
+ - **Compare source and current-target histories.** Compare authoritative metadata at the pinned source and current target, using the merge-base to attribute changes. Check lower and equal timestamps/versions, duplicate identities or timestamps, content-hash collisions/mismatches, missing journal entries or migration bodies, dependency gaps/order, and changed or removed already-published identities/content. Inspect relevant concurrently open PRs for competing identities/order when accessible; state that coverage limit when unavailable. Do not infer a conflict solely from similar filenames, or assume that a journal records a checksum the runner actually validates.
54
+ - **Separate branch history from applied database state.** A migration present in the target branch is not proof it was applied to any database. Label applied-state evidence with its environment, provenance and observation time when available; otherwise describe it as a hypothetical upgrade precondition, not a deployment fact. Never infer deployment state from merge order, CI success, or repository metadata alone.
55
+ - **Assess the late-arriving upgrade path.** Evaluate a database where a later migration has already been applied before an incoming earlier migration arrives, including incoming timestamps equal to the applied watermark. Use the runner's actual comparison, identity, dependency and guard rules to determine eligibility and consequences. A clean Git merge or successful fresh installation is not sufficient proof of upgrade safety. Use isolated fixtures/tests (with explicitly disposable state) or a reproducible code trace of the resolved runner and wrappers; never run a migration against a shared, staging, or production database. If isolation cannot be established, use code evidence and state the missing execution proof. Do not claim an isolated test was run when only tracing code.
56
+ - **Distinguish skip, SQL failure, and fail-closed refusal.** Neutral example: an incoming migration has `t_incoming <= t_applied`, while a later migration is already recorded as applied. These are conditional outcomes to investigate, not universal runner behavior:
57
+ - **Silent skipping:** if eligibility requires a timestamp strictly above the watermark, the incoming migration can remain unapplied while the runner reports success. Trace the resulting missing schema/data and affected application flow. **Error if not fixed: No** applies only when evidence establishes no error in that investigated flow; a later failed request is **Yes**, not a harmless skip.
58
+ - **SQL failure:** an individual-identity runner may still attempt the incoming migration, but a missing dependency or incompatible already-upgraded schema can make its SQL fail. Establish the reachable statement/error, transaction or partial-application consequences, and effect on startup/deployment; individual tracking alone does not prove safety. Use **Error if not fixed: Yes** only with demonstrated or code-established failure, otherwise **Unverified**.
59
+ - **Fail-closed refusal:** a reconciliation/startup guard may reject the history before SQL runs. Explain the guard condition, refusal/error, and resulting startup/deployment availability impact. Distinguish intentional protective behavior from a guard defect: a proven refusal of the required upgrade path is a release blocker even when the guard correctly prevents corruption. Never recommend bypassing it. Use **Error if not fixed: Yes** for an established refusal; otherwise **Unverified**.
60
+ - **Apply evidence-based verdicts.** Confirmed unsafe upgrade behavior (silent omission, SQL failure, or refusal preventing the required startup/deployment) is a **Release blocker** and requires **NOT SAFE TO MERGE**; cite the trigger, evidence and concrete unfixed consequence. Missing material runner/history/upgrade evidence requires **MERGE SAFETY UNVERIFIED** unless another established blocker or procedural gate already requires **NOT SAFE TO MERGE**. Put uncertain outcomes in Questions/Risks, not confirmed findings. Numbering differences alone are not a finding. Report which conclusions came from code versus isolated tests and what applied-state evidence remains unverified; a fully verified safe individual-identity upgrade is not a defect merely because its numbering differs.
61
+ - **No live repair or extra mutations.** The review remains read-only for code, records, and merge, with exactly one final PR summary comment by default (unless the user opts out). Do not auto-repair migration history, rewrite journals or applied ledgers, renumber applied migrations, change published identities/checksums, mark migrations applied, or bypass reconciliation/startup guards. Do not migrate shared databases, merge, deploy, or publish packages during review. Describe the smallest safe corrective approach for separately authorized implementation; do not execute it. Fold all migration findings and limitations into the single completed verdict/comment, never a separate mid-review comment.
62
+
49
63
  ## 4. Deliver a verdict
50
64
 
51
65
  - Use plain-text headings and numbered or bulleted lists only; **never use Markdown tables**. Keep the developer-facing response short, actionable, and easy to copy and paste. Do the full review and maintain internal ledgers, but do not print exhaustive inventories or checklists unless requested. Avoid repeating evidence across sections.