MCP Tools Reference

Daftar Tools

Oprex MCP mengekspos 224 tools yang mencakup seluruh siklus SDLC: projects, issues, requirements, test cases, milestones, releases, tickets, pipelines, notes, memories, webhooks, tags, deployments, dan lainnya.

Tool	Scope	Deskripsi
`oprex_list_projects`	read	List all projects in your Oprex tenant (GitLab-style). Returns id, name, slug, groupId, visibility.
`oprex_list_groups`	read	List all groups (namespaces) in your Oprex tenant. Returns id, name, path.
`oprex_resolve_project`	read	Resolve a GitLab-style path <group>/<project> to its group + project records. Use to turn a human path like 'indohrm/enterprise' into ids.
`oprex_list_issues`	read	List issues (bugs) — optionally scoped to a project and filtered by status/severity/text.
`oprex_list_bugs`	read	List bugs (issues) — alias for oprex_list_issues. Optionally scoped to a project and filtered by status/severity/text.
`oprex_get_issue`	read	Get a single issue (bug) by id, with full detail.
`oprex_create_issue`	write	Create a new issue (bug) — set severity, type, priority, labels (e.g. env:staging / env:rc), modules, and target/fixed versions.
`oprex_update_issue`	write	Update an issue (bug): status, severity, priority, type, title, description, labels, modules, target/fixed versions, or which project it belongs to.
`oprex_update_bug`	write	Update a bug (issue): status, severity, priority, type, title, description, labels, modules, target/fixed versions, or which project it belongs to.
`oprex_delete_bug`	write	Delete a bug (issue) permanently. Also removes any comments linked to it. This action cannot be undone.
`oprex_list_requirements`	read	List requirements — optionally scoped to a project and filtered by status/type/text.
`oprex_create_requirement`	write	Create a new requirement under a parent Specification. Pass specificationId to file it under an existing Specification in the same project (a Requirement belongs to one Specification). If none fits, create a Specification first (oprex_create_specification). When the workspace enforces parents (REQUIRE_SPEC_PARENT), specificationId is mandatory and the call is rejected without it.
`oprex_get_timeline`	read	Get a full timeline for a case (ticket, bug, or requirement) — events, SLA stats, AI vs manual work breakdown. Shows the complete lifecycle from creation to resolution.
`oprex_get_sla`	read	Get SLA status for a case (ticket, bug, requirement) — countdown timer, breach status, response/resolution deadlines. Supports 7 tiers: free, tim, komunitas, pro, bisnis, enterprise.
`oprex_get_audit_log`	read	Query the immutable audit log — ISO-27001 compliance tracking. Filter by actor, resource, action, time range. Tamper-proof record of all system events.
`oprex_get_audit_report`	read	Generate ISO-27001 compliance report — event summaries, outcome breakdown, incidents, data integrity status. Required for paid-tier SLA compliance.
`oprex_verify_integrity`	read	Run data integrity verification — checks timestamp ordering, missing actors, and anomaly detection. Part of ISO-27001 compliance.
`oprex_get_retention`	read	Get data retention policies per event type — ISO-27001 compliance. Shows retention days, auto-purge settings, and legal hold status.
`oprex_get_productivity`	read	Get AI vs manual work statistics for a time period. Shows productivity breakdown by event type, session, and daily timeline. Color coding: Green=AI, Orange=Manual.
`oprex_get_session_productivity`	read	Get productivity stats for a single AI session — event count, duration, AI vs manual breakdown.
`oprex_list_sla`	read	List all active cases with SLA status — filter by ok/warning/breached. Shows countdown timers and breach alerts.
`oprex_list_milestones`	read	List milestones — optionally scoped to a project and filtered by status.
`oprex_list_sprints`	read	List sprints — time-boxed iterations. Internally these are milestones tagged with module 'sprint', optionally scoped to a project.
`oprex_create_sprint`	write	Create a sprint (time-boxed iteration). Internally a milestone tagged with module 'sprint'.
`oprex_update_sprint`	write	Update a sprint: name, description, status, due date, or sort order.
`oprex_delete_sprint`	write	Delete a sprint by id.
`oprex_list_releases`	read	List releases (Staging → RC → Stable) — optionally scoped to a project.
`oprex_list_specifications`	read	List specifications (technical designs) — optionally scoped to a project.
`oprex_get_project`	read	Get a single project by id, with full detail (visibility, group, git repo, type).
`oprex_list_tickets`	read	List support/helpdesk tickets — optionally scoped to a project and filtered by status/text.
`oprex_list_test_cases`	read	List test cases (TestLink-style scenarios) — optionally scoped to a project.
`oprex_list_test_plans`	read	List test plans — optionally scoped to a project.
`oprex_list_test_runs`	read	List test runs — optionally scoped to a project or plan.
`oprex_list_pipelines`	read	List CI/CD pipelines and their runs — optionally scoped to a project.
`oprex_list_release_plans`	read	List release plans (roadmap of planned releases) — optionally scoped to a project.
`oprex_coverage_matrix`	read	Get the requirement→test coverage matrix for a project (traceability).
`oprex_list_notes`	read	List Notes (short sticky notes) — optionally scoped to a project, filtered by tag/pinned/text. Standalone notes have no sourceType attached.
`oprex_create_note`	write	Create a standalone Note (short sticky note, max ~20k chars).
`oprex_list_memories`	read	List Memory documents (long-form markdown imported from AI vibe-coding sessions, e.g. Claude Code). scope='personal' is private to the caller; scope='project' is shared with the project team.
`oprex_get_memory`	read	Get a single Memory document by id, including its full markdown body.
`oprex_create_memory`	write	Create a Memory document — save a markdown file/summary from a vibe-coding session (Claude Code etc.) into Oprex so a team can read/discuss it. scope='project' requires projectId.
`oprex_create_ticket`	write	Create a support/helpdesk ticket.
`oprex_update_ticket`	write	Update a ticket: status, priority, severity, subject, body, or assignee.
`oprex_delete_ticket`	write	Delete a ticket permanently. Also removes any replies linked to it. This action cannot be undone.
`oprex_update_requirement`	write	Update a requirement: status, title, description, acceptance criteria, type, or priority. Note: setting status to 'approved' automatically creates a test case stub prefixed with '[Auto]' for traceability.
`oprex_delete_requirement`	write	Delete a requirement permanently. This action cannot be undone.
`oprex_add_comment`	write	Add a comment to any Oprex artifact (bug, ticket, requirement, test_case, test_plan, test_run, milestone, release, pipeline, project, or specification) — the thread every artifact detail page shows.
`oprex_list_comments`	read	List the comment thread on any Oprex artifact.
`oprex_trigger_autopilot`	write	Manually trigger Autopilot (AI auto-repair) on a bug: diagnoses the root cause, drafts a fix, opens a PR for human review. NOTE: restricted to Kinexa staff accounts — a tenant/customer API key will get a 403 here even if it works for other tools.
`oprex_update_memory`	write	Update an existing Memory document — replace body/title/tags, RE-PARENT it to a different project (scope='project' + projectId), switch to personal, or archive it. Use this to fix a memory whose project parent is wrong or stale.
`oprex_delete_memory`	write	Archive (soft-delete) a Memory document by id. The memory is retained for audit but hidden from default lists.
`oprex_list_graph_edges`	read	List entity-relation edges for the memory graph. Filter by subject, predicate, or object. Returns active edges by default; pass status='deleted' to inspect soft-deleted edges.
`oprex_get_graph_edge`	read	Get a single graph edge by id.
`oprex_create_graph_edge`	write	Add a relation between two entities in the memory graph. Optional sourceFactId links the edge to a memory fact so superseding that fact can be audited later.
`oprex_delete_graph_edge`	write	Soft-delete a graph edge (status -> deleted). The row is retained for audit.
`oprex_list_backlogs`	read	List TODO/backlog items for the authenticated user or project. Use this to resume work, track what AI was doing, and maintain context across sessions. Filter by state, project, or source type.
`oprex_list_todos`	read	List TODO items — alias for oprex_list_backlogs. Optionally scoped to a project, session, or source type.
`oprex_create_backlog`	write	Create a TODO/backlog item — a unit of work that AI or human must do next. Every backlog item carries an audit trail (who created it, from which source artifact) so work is never lost and can be resumed after context compaction.
`oprex_update_backlog`	write	Update a TODO/backlog item: state (pending/done/snoozed/dismissed), priority, due date, or description. Use this to mark work complete when a release/test run finishes, or to reschedule when blocked.
`oprex_delete_todo`	write	Delete a TODO permanently. This action cannot be undone.
`oprex_update_memory_on_release`	write	Standard SDLC operation: when a release is shipped, update the project's Memory document with the release summary, linked requirements, test results, and any issues resolved. This keeps the Memory in sync with reality automatically — no manual edit required. Use this after calling oprex_ship_release_plan or oprex_create_release.
`oprex_update_memory_on_test_run`	write	Standard SDLC operation: when a test run finishes, update the project's Memory document with the test summary, coverage status, and linked requirements. Use this to keep traceability current without manual copy-paste.
`oprex_create_test_case`	write	Create a test case (TestLink-style scenario). Note: requirementIds may only reference APPROVED requirements — linking a draft one is rejected.
`oprex_update_test_case`	write	Update a test case: title, summary, preconditions, steps, expected result, priority, execution type, linked requirements, tags — or archive it with isActive='false'.
`oprex_delete_test_case`	write	Delete a test case permanently. Also removes any test results linked to this case. This action cannot be undone.
`oprex_create_test_plan`	write	Create a test plan. A test plan groups test cases under a name/version so you can run them as a batch.
`oprex_update_test_plan`	write	Update a test plan: name, description, version, status, caseIds, or milestoneId.
`oprex_delete_test_plan`	write	Delete a test plan permanently. This also affects any test runs linked to it.
`oprex_create_test_run`	write	Create a test run (execution instance) against an existing test plan. A run records build version, environment, and per-case results.
`oprex_update_test_run`	write	Update a test run: name, build version, environment, or status.
`oprex_record_test_result`	write	Record the result of a single test case within a test run. Sets pass/fail/blocked/skip status, optional notes, and can link a bug. Auto-promotes the run from 'planned' to 'in_progress' on first execution.
`oprex_delete_test_run`	write	Delete a test run permanently. This also removes any case results linked to it.
`oprex_finish_test_run`	write	Finish a test run. Marks the run as completed and records the finish timestamp.
`oprex_create_release_plan`	write	Create a release plan (a planned release you then fill with scope — requirements, bugs, test cases — before shipping it as a Release with oprex_ship_release_plan).
`oprex_update_release_plan`	write	Update a release plan: name, description, version, status, target date, or milestone. To ship one, use oprex_ship_release_plan — that publishes the Release notes too, which setting a status here does not.
`oprex_ship_release_plan`	write	Ship a release plan: regenerates its release notes from the current scope, publishes them as an internal Release, and marks the plan shipped. Fails if the plan was already shipped.
`oprex_release_plan_add_scope`	write	Add a bug, requirement, or test case to a release plan's scope. Use this to link work items to a planned release before shipping it.
`oprex_create_release`	write	Publish a Release (release notes) — version, title, and notes grouped into the standard changelog categories. Created as INTERNAL: visible to your tenant (changelog + What's New popup), never on the public oprex.id changelog — announcing publicly stays with Kinexa staff.
`oprex_update_release`	write	Update a published Release: version, title, severity, notes, milestone, or project. Any note category you send REPLACES the whole notes block, so resend every category you want to keep — read the release first with oprex_list_releases.
`oprex_get_release_report`	read	Generate a structured release report for a release: summary stats, charts (release cadence, bug trend, test coverage, workload, severity distribution), scope summary, deployment fingerprint, activity log, requirements coverage, and bug resolution time. Data is read-only and generated from existing Oprex artifacts.
`oprex_delete_release`	write	Delete a release permanently. This action cannot be undone.
`oprex_update_milestone`	write	Update a milestone: title, description, status, due date, sort order, or the modules it covers.
`oprex_update_specification`	write	Update a specification (technical design): title, summary, body, version, status, or modules. Each body change is kept as a revision, so the diff history stays intact.
`oprex_delete_specification`	write	Delete a specification permanently. This action cannot be undone.
`oprex_update_note`	write	Update a Note (sticky note): title, body, color, or pinned state.
`oprex_delete_note`	write	Delete a note permanently. This action cannot be undone.
`oprex_create_milestone`	write	Create a milestone. Nest it under another one with parentId — milestones form a tree, which is what the roadmap view renders.
`oprex_create_specification`	write	Create a specification (technical design). The body is markdown and every later change is kept as a revision, so the spec carries its own diff history.
`oprex_create_group`	write	Create a group (namespace, GitLab-style). Needs owner/maintainer on the tenant, and a tenant-wide API key — a key scoped to specific Groups/Projects is refused, since creating a namespace is a tenant-level act.
`oprex_update_group`	write	Update a group: name, description, confidential flag, or client codename.
`oprex_create_project`	write	Create a project inside a group. Needs owner/maintainer plus a tenant-wide API key (a Group/Project-scoped key is refused), and counts against your tier's project quota. A confidential group forces its projects private whatever visibility you ask for.
`oprex_update_project`	write	Update a project's name, description, or type. Slug, group, visibility and the confidential flag are NOT editable here: they also define the backing git repo, and Oprex has no way to rename or re-permission that repo in the same step — changing them on this side alone would make the two disagree.
`oprex_trigger_pipeline`	write	Manually run a CI/CD workflow on a project's git repo and record the pipeline so it shows up in Oprex right away. repoFull must be a repo already linked to one of your projects (see forgejoRepo on oprex_get_project).
`oprex_recall_memory`	read	Search remembered FACTS (atoms), not documents — ranked full-text over the statement itself, so a fact buried mid-paragraph is still found (oprex_list_memories only matches a document's title and source label). Call this BEFORE oprex_remember: the fact ids you get back are what you pass to `supersedes` when something has changed.
`oprex_remember`	write	Write what you learned as facts, in one transaction: add / update / delete / noop per operation. Recall first, then reconcile against what is already there — a fact that replaces an older one MUST name it in `supersedes`, which marks the old fact superseded (kept, linked, auditable) instead of leaving two contradictory facts alive. Use noop when a fact is already recorded and still true; that raises its confirmation count rather than duplicating it. The response returns `coexisting`: active facts on the same subject you did NOT supersede — check them, that is where a silent contradiction would hide.
`oprex_memory_fact_history`	read	Get one fact plus every change it went through — each revision holds the state BEFORE that change, with who changed it and why. Answers 'when did we stop believing this, and what replaced it'.
`oprex_search`	read	One ranked full-text search across issues, tickets, requirements, specifications, test cases, milestones and notes. Use this FIRST when the question is 'has this already been written down somewhere' — it replaces calling seven list tools and reading seven full payloads, and returns only a short snippet per hit, so follow up with the matching get tool for full detail. Indexed under both Indonesian and English stemmers, so 'pendaftaran' also matches 'mendaftar'. Terms are OR-ed and ranked, so a long question still returns its best matches instead of nothing. Memory is deliberately NOT searched here because its access rules are per-user — use oprex_recall_memory for facts and oprex_list_memories for documents.
`oprex_list_kb_categories`	read	List Knowledge Base categories in your tenant. Returns id, slug, name, description, icon, sortOrder.
`oprex_list_kb_articles`	read	List Knowledge Base articles — optionally filtered by category, language, visibility, or free text. Returns id, slug, title, excerpt, categoryId, visibility, lang, tags, readingMinutes.
`oprex_get_kb_article`	read	Get a single Knowledge Base article by slug, with full body. Pass lang to pick a translation.
`oprex_create_kb_article`	write	Create a Knowledge Base article. slug, title and body are required; category, visibility, lang, tags, relatedIds, isPublished, status, changeSummary, autoGeneratedFrom are optional.
`oprex_update_kb_article`	write	Update a Knowledge Base article by id — creates a new version. changeSummary is required.
`oprex_delete_kb_article`	write	Soft-delete a Knowledge Base article by id.
`oprex_list_kb_versions`	read	List version history for a KB article.
`oprex_get_kb_version`	read	Get a single version of a KB article.
`oprex_compare_kb_versions`	read	Compare two versions of a KB article.
`oprex_restore_kb_version`	write	Restore a previous KB version as a new version.
`oprex_submit_kb_review`	write	Submit a KB article for review.
`oprex_approve_kb_article`	write	Approve a KB article for publication.
`oprex_reject_kb_article`	write	Reject a KB article back to draft.
`oprex_schedule_kb_article`	write	Schedule a KB article for future publication.
`oprex_archive_kb_article`	write	Archive a KB article with optional deprecation note and superseded-by link.
`oprex_auto_generate_kb_from_artifact`	write	Create a draft KB article from an existing artifact (requirement, bug, memory, release).
`oprex_list_webhooks`	read	List webhooks in your tenant. The secret is never returned — only a short preview.
`oprex_webhook_events`	read	List the event names a webhook can subscribe to (e.g. 'bug.created', 'release.shipped'), with descriptions. Use to pick the `events` argument when creating a webhook.
`oprex_get_webhook`	read	Get a single webhook by id, with its subscribed events and active flag.
`oprex_webhook_deliveries`	read	List recent delivery attempts for a webhook (status, timing, response) — useful for debugging why an integration didn't fire.
`oprex_create_webhook`	write	Create a webhook. url is required; description and an `events` allow-list (from oprex_webhook_events) are optional. The secret is returned ONLY on create.
`oprex_update_webhook`	write	Update a webhook: change url, description, subscribed events, or toggle isActive.
`oprex_delete_webhook`	write	Delete a webhook by id.
`oprex_list_tags`	read	List tag definitions in your tenant, optionally filtered by text or project. Returns id, name, color, description, usageCount.
`oprex_create_tag`	write	Create a tag definition. name is required; color (#RRGGBB) and description are optional.
`oprex_update_tag`	write	Update a tag definition: name, color, and/or description.
`oprex_delete_tag`	write	Delete a tag definition by id. Also removes it from every entity it was applied to.
`oprex_entity_tags`	read	List the tags currently applied to any artifact (a bug, requirement, test case, spec, release, project, KB article, …).
`oprex_tag_entity`	write	Tag an artifact. Pass an existing tagId, OR a name (the tag is created on the fly if it doesn't exist yet).
`oprex_untag_entity`	write	Remove a tag from an artifact.
`oprex_tag_usage`	read	Show where a tag is used — which artifacts it's applied to, grouped by kind, with totals. Use to find everything carrying a given label.
`oprex_get_activity`	read	Get the change history of any artifact — who did what and when (created, status changes, comments, attachments). sourceType is the artifact kind, sourceId its id.
`oprex_get_tool_telemetry`	read	Get usage statistics for MCP tools in the current session: total calls, last called time, and error count per tool. Useful for understanding tool usage patterns.
`oprex_generate_dev_report`	read	Generate a Markdown + JSON report summarizing developer contributions: commits, artifacts (requirements, bugs, tests, releases), deployments, PRs, and issues. Pulls data from Oprex MCP, Git log, GitHub API, and CI/CD.
`oprex_move_artifact`	write	Move a bug, requirement, specification, test case, note, ticket, or milestone to another project. Actor must be a member of both source and target projects. Activity is logged.
`oprex_get_daily_summary`	read	Get a quick summary of today's activity: open bugs, pending tickets, recent requirements, and upcoming milestones. Use this for daily standup preparation.
`oprex_get_sso_status`	read	Check if SAML/SSO is enabled for the tenant. Returns enabled status and button label.
`oprex_get_saml_settings`	read	Get the current SAML/SSO configuration for the tenant. Includes IdP/SP entity IDs, URLs, and certificate info.
`oprex_update_saml_settings`	write	Create or update SAML/SSO configuration for the tenant. Set IdP entity ID, SSO/SLO URLs, certificates, and enable/disable SSO.
`oprex_list_deployments`	read	List deployment runs (newest first). Pass env to scope to one environment, status:'live' to get what's currently deployed there. Each run carries commitSha, fingerprint and fileCount once the target host has reported them.
`oprex_get_environment`	read	Get one deploy environment (app × env) with its LIVE snapshot — currentVersion, currentImageRef, and the mirrored fingerprint + fileCount of what's actually deployed right now.
`oprex_compare_environments`	read	Diff two deploy environments at file granularity — answers 'are Nila and Roa identical?'. Returns same plus which files are only-in-A / only-in-B / changed. Uses each env's LIVE fingerprint (from the target host's scan).
`oprex_report_deployment_manifest`	write	Stamp the content fingerprint + file count measured on the deploy target host for whatever is live in an environment. Run scripts/scan-deploy-manifest.ts on the host, then send its fingerprint + fileCount here so Release Readiness can compare environments.
`oprex_list_deploy_presets`	read	List deployment config presets for the current tenant's applications. Each preset stores Coolify connection details (API URL, server UUID, project UUID, env) so sessions can discover how to deploy without re-typing credentials. Never returns the raw API token.
`oprex_get_deploy_preset`	read	Get a single deployment config preset by id, including Coolify connection details. Never returns the raw API token — only whether it is configured.
`oprex_get_auto_behaviour`	read	Return the current user's effective MCP auto-behaviour preferences: mode (manual / auto-read / auto-save-light / auto-save-full), repo→project mapping, and artifact numbering format. Use this at session start to decide whether to auto-load Oprex context.
`oprex_session_start`	write	Call this at the start of a vibe-coding session. Creates or resumes a session, loads context for the selected project(s), and returns session info + artifacts. If projectIds is not provided, returns ask:true so the user can select projects.
`oprex_session_end`	write	Call this at the end of a vibe-coding session. Creates a memory summarizing the session, links all artifacts created during the session, and marks the session as completed.
`oprex_list_my_work`	read	Get a unified view of all work assigned to the current user — pending todos, open bugs, and assigned tickets. Returns grouped by artifact type.
`oprex_update_project_cascade_mode`	write	Update the cascade_mode setting for a project. Controls how requirement backupdates cascade to linked tickets/bugs. Modes: off (no cascade), suggest (create pending suggestions), auto (apply automatically).
`oprex_list_notifications`	read	List notifications for the current user, optionally filtered by read/unread status.
`oprex_get_notification`	read	Get a single notification by id.
`oprex_mark_notification_read`	write	Mark a single notification as read.
`oprex_mark_all_notifications_read`	write	Mark all notifications for the current user as read.
`oprex_notification_unread_count`	read	Get the count of unread notifications for the current user.
`oprex_traceability_matrix`	read	Get a project-level traceability matrix: requirements → test cases → bugs → releases in one call.
`oprex_list_requirement_backupdates`	read	List pending requirement back-update suggestions (test run / bug / ticket → requirement status change).
`oprex_apply_requirement_backupdate`	write	Apply a pending requirement back-update suggestion.
`oprex_dismiss_requirement_backupdate`	write	Dismiss a pending requirement back-update suggestion without applying it.
`oprex_list_documents`	read	List documents — optionally scoped to a project or group, with optional full-text search.
`oprex_get_document`	read	Get a single document by id.
`oprex_create_document`	write	Create a document (wiki page). Supports markdown body, tags, and optional project/group scoping.
`oprex_update_document`	write	Update a document: title, body, slug, tags, or published state.
`oprex_delete_document`	write	Delete a document by id.
`oprex_list_templates`	read	List compliance/industry templates — optional category filter (e.g. 'compliance', 'medical', 'automotive').
`oprex_get_template`	read	Get a single template by id.
`oprex_create_template`	write	Create a custom compliance/industry template. The definition is a JSON blob describing artifacts to scaffold.
`oprex_apply_template`	write	Apply a template to a project. The template definition JSON must include 'requirements', 'testCases', and/or 'milestones' arrays to scaffold.
`oprex_bulk_update_issues`	write	Update multiple issues in one call. Each item needs the bug id plus the fields to patch. Max 100 items per call.
`oprex_bulk_create_requirements`	write	Create multiple requirements in one call. Accepts an array of requirement payloads. Max 100 items per call.
`oprex_export_project`	read	Export all artifacts of a project as a single JSON document — requirements, test cases, bugs, milestones, specifications, and releases. Use for backup or migration.
`oprex_import_project`	write	Import artifacts into a project from a JSON payload. Creates new requirements, test cases, bugs, milestones, and specifications. Existing artifacts are not modified.
`oprex_import_jira`	write	One-click import from Jira JSON export. Transforms Jira issues (bugs, stories, tasks) into Oprex requirements, bugs, and milestones. Accepts Jira's native JSON export format. Maps issue types, priorities, statuses, fix versions, and labels automatically.
`oprex_import_linear`	write	One-click import from Linear JSON export. Transforms Linear issues into Oprex requirements, bugs, and milestones. Accepts Linear's GraphQL export format. Maps priorities, states, labels, projects, and due dates automatically.
`oprex_list_features`	read	List all available features that can be scoped to API keys. Returns feature key, name, category, and description.
`oprex_get_api_key_features`	read	Get the feature scopes for an API key. Returns list of feature keys this key is restricted to. Empty list means full access.
`oprex_create_session`	write	Register a new AI/terminal session (e.g. Kilo, Claude Code, Gemini CLI). Returns the session id — store it and send heartbeat via oprex_touch_session.
`oprex_list_sessions`	read	List AI/terminal sessions for this tenant. Filter by status, type, model, provider, or search alias.
`oprex_get_session`	read	Get a single AI/terminal session by id.
`oprex_update_session`	write	Update an AI/terminal session — alias, model, provider, or status.
`oprex_delete_session`	write	Soft-delete an AI/terminal session. Can be restored with oprex_restore_session.
`oprex_archive_session`	write	Archive an AI/terminal session (status → archived). Archived sessions are hidden from the default active list but preserved for history.
`oprex_restore_session`	write	Restore an archived or deleted AI/terminal session back to active.
`oprex_touch_session`	write	Update the last activity timestamp of an AI/terminal session. Call this periodically (e.g. every 30s) from your AI tool to keep the session alive.
`oprex_session_stats`	read	Get session statistics for this tenant: total, active, archived, by type, by model.
`oprex_session_split`	write	Split a multi-project session into separate sessions per project. Supports auto-detection of project ownership for unassigned artifacts using keyword matching.
`oprex_session_merge`	write	Merge multiple sessions into a single session. Source sessions are archived after merge.
`oprex_session_import`	write	Import a vibe-coding session from an external AI tool (Claude Code, Kilo, Cursor, Windsurf, Cline, Aider, Codex, OpenRouter, GitHub Copilot, Replit Agent, Bolt.new, v0, Lovable, Cody, Tabnine, Continue.dev, Amazon Q, Gemini CLI, JetBrains AI, Supermaven, or generic JSON). Creates a memory document from the imported content.
`oprex_session_import_process`	write	Process a pending session import — parses the raw data and creates artifacts (memories, notes) from the imported session content.
`oprex_list_session_imports`	read	List session imports with optional filtering by status and source tool.
`oprex_list_prompts`	read	Browse the enterprise prompt library — filterable by category, target tool, language, framework, and free text search.
`oprex_get_prompt`	read	Get a single prompt template by id with full content and placeholder definitions.
`oprex_create_prompt`	write	Create a new prompt template in the library. Use {{placeholder}} syntax in content for dynamic values.
`oprex_generate_prompt`	write	Generate a ready-to-use prompt by filling template placeholders with provided context values.
`oprex_apply_prompt`	write	Apply a prompt template to the current session — generates the prompt with session context and links it to the session.
`oprex_ai_chat`	read	Send a message to the configured AI model and get a response. Uses tenant AI settings from master.kinexa.id.
`oprex_autopilot_analyze`	read	Analyze a project and generate prioritized task list from open bugs, pending backlogs, and draft requirements.
`oprex_orchestrate`	write	Run full orchestration: analyze all tasks, execute them in sequence, return results.
`oprex_orchestrate_multi`	write	Run orchestration across multiple projects simultaneously with worker pool.
`oprex_detect_project`	read	Auto-detect which project is being discussed. Uses LLM + keyword matching.
`oprex_extension_list`	read	List all configured external tool integrations (Jira, Linear, Asana, etc.).
`oprex_deploy`	write	Deploy a project to an environment (dev, staging, rc, production). Supports Coolify, ArgoCD, Jenkins, GitHub Actions, GitLab CI.
`oprex_deploy_status`	read	Get deployment status by ID or list all deployments for a project.
`oprex_deploy_pipeline`	read	Get deployment status across all environments (dev, staging, rc, production) for a project.
`oprex_deploy_approve`	write	Approve a pending deployment to proceed (required for staging, rc, production).
`oprex_deploy_rollback`	write	Rollback a deployment to previous state.
`oprex_role_prompt`	read	Get the system prompt for a specific role (developer, qc, collaborator, admin). Shows what AI sees in each role.
`oprex_role_list`	read	List all available roles with their allowed actions and descriptions.
`oprex_event_publish`	write	Publish an event to the Oprex event bus. Triggers auto-action rules. Events: issue.created, commit.pushed, deployment.started, etc.
`oprex_auto_action_list`	read	List all auto-action rules for the tenant. Rules trigger actions when events fire.
`oprex_auto_action_create`	write	Create a new auto-action rule. When the trigger event fires and conditions match, actions execute automatically.
`oprex_ai_model_get`	read	Get AI model configuration for the tenant. Returns provider, models, and limits.
`oprex_ai_model_save`	write	Save AI model configuration for the tenant. Providers: longcat, anthropic, openai, google, custom.
`oprex_ai_model_providers`	read	List all supported AI providers with their default model IDs.
`oprex_approval_list`	read	List approval requests. Filter by status (pending/approved/rejected) and type (deployment/autopilot/release).
`oprex_approval_approve`	write	Approve a pending request. Executes the approved action (deployment, autopilot, release).
`oprex_approval_reject`	write	Reject a pending request with a reason.
`oprex_approval_stats`	read	Get approval statistics: pending, approved, rejected counts.
`oprex_onboarding_progress`	read	Get onboarding progress for the tenant. Shows completed steps and total.
`oprex_onboarding_complete`	write	Mark an onboarding step as completed.
`oprex_filing_list`	read	List enterprise filing documents. Filter by category (legal, compliance, contract, policy).
`oprex_filing_create`	write	Create an enterprise filing document. Supports encryption for sensitive documents.
`oprex_list_skills`	read	List available skills that AI can load. Returns skills filtered by category, tool, tags, or search query.
`oprex_load_skill`	read	Load a skill by ID or code. Returns the full skill content that AI can use to execute tasks.
`oprex_create_skill`	write	Create a new skill. Skills are structured prompt/template documents that AI can load and execute.
`oprex_update_skill`	write	Update an existing skill by ID.
`oprex_delete_skill`	write	Delete a skill by ID.

Perkecil Tool Surface dengan Toolsets

Server Oprex MCP mengekspos 79 tool. Jika Anda hanya butuh sebagian — misalnya hanya issues dan requirements — Anda bisa membatasi daftar tool yang dimuat ke dalam konteks model. Ini menghemat token definisi tool pada setiap sesi.

Cara penggunaan:

Via JSON-RPC params pada tools/list: toolsets: "issues,requirements"
Via HTTP header: X-MCP-Toolsets: issues,requirements
Via query string: ?toolsets=issues,requirements

Kategori toolsets yang tersedia:

Toolset	Isinya	Contoh penggunaan
`all`	Semua 79 tool	Default, tanpa filter
`readonly`	Semua tool read-only	Hanya-baca, tanpa write
`write`	Semua tool write	Hanya-tulis, tanpa read
`issues`	`list/get/create/update`	Bug triage
`requirements`	`list/create/update`	Product management
`tests`	`list/create/update + coverage`	QA workflows
`releases`	`releases + release plans`	Release management
`deploy`	`deployments + environments`	DevOps / SRE
`memories`	`docs + facts + graph`	Knowledge / AI memory
`search`	`oprex_search`	Cross-artifact lookup only

Kategori lain yang tersedia: projects, milestones, specifications, tickets, pipelines, notes, comments, webhooks, tags, activity, kb, backlog, autopilot, auto-behaviour. Gabungkan dengan koma, misalnya issues,tests,releases.

Structured Responses

Tool yang mengembalikan objek JSON kini menyertakan structuredContent bersamaan dengan teks, sehingga klien yang mendukungnya bisa parsing langsung tanpa escaped JSON.

Ini mengurangi overhead parsing dan memudahkan programmatic consumption.