Returns the authenticated Mantis user (id, name, access level, accessible projects). Use it to verify the API token works.

Lists all projects accessible to the token owner, including nested versions, categories and custom field definitions.

Returns a single project by id with its versions, categories and custom fields.

Lists the members of a project (id, login name, access level). Use it to find valid handler names for assignment.

Returns the common Mantis enumerations (status, priority, severity, resolution, reproducibility, etc.) as {id,name,label} lists. This instance is localized (Spanish names) — use it to map issue status/priority codes and to learn valid names for status changes.

Reads arbitrary Mantis configuration options exposed to the REST API, e.g. ["status_enum_string","bug_resolved_status_threshold"].

Lists stored filters, optionally scoped to a project. Pass a filter id to mantis_list_issues to run it.

Returns one issue by id with every field: description, steps to reproduce, notes, attachments, relationships, custom fields and history. Notes are always included.

Fetch several issues by id in one call (max 5 concurrent). Missing/inaccessible ids return null at their position instead of failing the whole call. Includes requested/found/failed counters.

List issues (newest id first). Returns a LIGHTWEIGHT summary per issue by default (id, summary, status, handler, reporter, priority, dates…) — call mantis_get_issue for notes/attachments/history. Filter server-side by project_id / stored filter_id. The status, assigned_to, reporter_id and date filters are applied client-side (the REST API lacks them); when active, multiple pages are scanned (up to 500 issues). Use status="open" for all not-yet-resolved issues. Omit project_id to span all projects.

Downloads an attachment of an issue. Image attachments are returned as a viewable image; other files as base64 with metadata. Find file ids in an issue's "attachments" array.

Partial update of an issue (PATCH). The "fields" object accepts any of:

• summary, description, steps_to_reproduce, additional_information (strings)

• status, resolution, priority, severity, reproducibility: { name } (localized or canonical) or { id }

• handler: { name: "" } or { id }

• category, version, target_version, fixed_in_version, view_state: { name }

• custom_fields: [{ field: {id|name}, value: "" }]

When resolving an issue set BOTH status and resolution (or use mantis_resolve_issue). Use "note" to append a comment (e.g. the reason for the change) in the same call. Use dry_run to preview the payload.

Marks an issue as resolved by setting BOTH status and resolution in one call (avoids leaving resolution as "open"). Defaults: status="resuelta", resolution="fixed" — override with names valid on this instance (see mantis_get_enums). Optionally append a note.

Assigns an issue to a handler (by login name or user id) and optionally changes its status in the same call. Optionally append a note.

Adds a note (comment) to an issue. Use private=true for an internal note visible only to developers/managers. To change a field and comment in one call, use mantis_update_issue with its "note" parameter instead.

Links two issues. type is one of: related_to, duplicate_of, has_duplicate, depends_on (parent_of), blocks (child_of) — or a numeric type id.

Removes a relationship from an issue by relationship id.

Search issues by natural-language meaning (not keywords), useful to find similar/duplicate incidents. The index must be built first with mantis_rebuild_search_index. Without "select" only id+score are returned.

Fill level of the semantic index: indexed vs. total issues, plus last sync timestamp.

Builds/updates the semantic search index from Mantis issues. Use full=true to clear and rebuild from scratch. First run downloads the embedding model (may take a minute).