Archipelago
Server Details
A living map of all your work that keeps you, your team, and your agents on the same page
The last two health checks of this connector failed. If this server requires authentication, missing or expired test credentials may be the cause. A test profile lets Glama authenticate for health checks and discover tools; it is separate from your personal connections.
If you are the author, claim ownership, then add or update a test profile under Admin → Test Profile.
- Status
- Unhealthy
- Uptime
- 0.0% over the last 24 hours
- OAuth
- Works in Glama
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- Repository
- AlexReiner98/archipelago
- GitHub Stars
- 1
- Server Listing
- Archipelago
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 10 tools
Several tools clearly overlap: change claims to be 'the only way to change the map' while revise also 'changes, corrects or undoes' and undo reverses change sets, leaving three mutation entry points with fuzzy boundaries. Similarly capture ('a record of what happened') vs report ('say that something happened') vs the session actions split across both, and view rendering is spread over query.look, draw, and report.show.
Every tool is a single lowercase verb in the same convention (ask, assign, capture, change, draw, query, report, revise, settle, undo), which is predictable. The only weakness is that the verbs carry no object, so close synonyms (change/revise/undo, capture/report) rely entirely on descriptions to differentiate.
Ten top-level tools is a reasonable number for a map/coordination server. However, each tool is really a dispatcher over 5-30 embedded actions (change alone has eight ops, query lists dozens of reads/views), so the modest count masks a surface well over a hundred operations and makes the set heavier than it looks.
The lifecycle looks well covered: reads (query), mutations with create/update/link/archive/forget/restore/merge (change), reversal (undo/revise), decision closure (settle), resource claims (assign), notifications (ask), sessions/reporting, backup, and rendering. Minor gaps remain around identity/permissions and cross-tool policy visibility, but core workflows have no dead ends.
Available Tools
10 toolsaskAskCInspect
Put a question to someone and wait for the answer. Call with the action and its params. Its actions: notify (title, body?, level?, about?); publish_ask (within, reason?); share_ask (id, reason?). query actions says what each does.
Inputs: actor, action, params.
assignAssignBInspect
Give work to someone, inside the map or outside it: an owner, a person waited on, a party asked to act. Call with the action and its params. Its actions: claim (resource, holder, for?, reason). query actions says what each does.
Inputs: actor, action, params.
captureCaptureBInspect
Put something on the map: an item and its links, or a record of what happened. Call with the action and its params. Its actions: session_start (title?, task?, surface?, cwd?, remote?, session_id?, branch?, parent?, source?); picture_put (result, data, content_type); backup (none); import_audit (connection); menu.save_view (id?, title, filter, walk?, limit?, time?, sessions?, style?, preset?). query actions says what each does.
Inputs: actor, action, params.
changeChangeADestructiveInspect
The only way to change the map: a list of ops with a reason. A create’s ref lets later ops link to it, and a title or note in the same change names what it makes as {ref}. Give the source the facts came from. Change part of a note with append or replace, of code with edit, not the whole. An update of only the id says it is still true. A plan or research are items with a view, not one note: a finding is evidence, an item that bears_on the question it helps settle. Checked whole, then applied or held for a human, by policy. With holds off it applies at once: public reaches anyone, code runs for all who open it, forget is for good; stakes say what it reached. A refused change is kept a few minutes under a name the refusal gives: send retry and fixes (the wrong ops, by number). A settled decision is put to the owner later (attention, to_confirm). Each op is { op, ...its fields }: create (ref?, type, title, body?, status?, owner?, visibility?, confidence?, props?, source?); update (id, title?, body?, append?, replace?, edit?, status?, owner?, visibility?, confidence?, props?, as_of?); link (type, from, to, props?, confidence?, source?); unlink (relation); archive (id); forget (id, reasons?); restore (id); merge (keep, drop); define_type (type). query actions with action change says each field.
Inputs: key, ops, run, hold, +7 more.
drawDrawARead-onlyInspect
Draw a view in this chat, for the person reading it, in a client that draws MCP Apps (Claude chat). Elsewhere nothing is drawn: report show opens it on their canvas, query look gives you the picture.
Inputs: actor, query, style, entity.
queryReadARead-onlyInspect
Read the map: name what to read, with its params. With no name, each name with what it answers. actions says what one takes, whole: params read (a read or view, such as flat), action (or change, for its ops) or verb. Answers drop empty fields and cut long lists (params.full: true for all). The reads and their params: schema (type?, full?, island?, reads?); search (q?, type?, status?, owner?, open?, include_archived?, islands?, limit?, full?); get (id, bare?, history?, full?); events (since?, target?, actor?, by?, before?, origin?, limit?, full?); proposals (status?, full?); picture (result?, look?); watch (after?, wait_seconds?); packs (none); acts (none); backup_download (none); look (islands?, entity?, query?, lens?, style?, width?, height?, scheme?, fresh?, keep?, published?, press?, marks?, at?, tries?); seeing (run?); document (id, drawn?, page?); publish_plan (within); published (id). Every other name is a view: flat, changed_since, spent, usage, notifications, sessions, session_measures, story, turns, parts, saved_views, style, instructions, hubs, frontier, attention, graph, reading, shared, misses, style_presets, layouts, brief. flat is the query: filter { find, only, not }, walk { over (part_of< down, part_of> up), steps }, limit.
Inputs: name, actor, params.
reportReportCInspect
Say that something happened or is so: progress, a result, a payment, a check that it still holds. show answers the owner with a view. Call with the action and its params. Its actions: session_update (run, status?, note?, task?, branch?, title?, parent?); session_end (run, outcome?, reason?, summary?, still_true?, cannot_tell?, cannot_tell_rest?); tell (run, said, about?, links?, replaces?, finding?); run_report (repo, title, ok?, tokens?, cost_usd?, turns?, link?, summary?, surface?); rechecked (repo, summary?); show (islands?, entity?, query?, lens?, style?, screen?, url?, note?, watch?, look?, width?, height?, scheme?, fresh?, pane?, run?); preview (item, field, value?, discard?); check_connections (none); health (heartbeat?, monitor?, msg?, live?, idle?, said?); hook (event, session_id, cwd?, prompt?, source?, stop_hook_active?). query actions says what each does.
Inputs: actor, action, params.
reviseReviseADestructiveInspect
Change, correct or undo what is on the map. Call with the action and its params. Its actions: release (resource, holder?, reason); withdraw (proposal, reason?); publish (id, authority?); publish_all (within, authority?); pack_set (pack, on); undo (changeset); menu.drop_view (id); styles.use (preset, use?); styles.delete_layout (layout). query actions says what each does.
Inputs: actor, action, params.
settleSettleCInspect
Close a question by someone's word or by evidence: a decision made, a hypothesis borne out. Call with the action and its params. Its actions: confirmed (upto, basis). query actions says what each does.
Inputs: actor, action, params.
undoUndoADestructiveInspect
Reverse one change set (its id is on every event it logged) by applying the opposite as a new change: created items are archived, edits put back, links removed or restored. An edit changed again since is left alone and reported. Subject to the same policy as any change.
Inputs: changeset (required), actor.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Organize work with your team and agents. Manage tasks, projects and documents.
Dashboards, trackers and docs your agents build and keep current. Your team reads them by link.
Keep teams & agents coordinated automatically
The plan your AI agents work from. Project management and customer feedback for agent teams.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceEnables agents to build and maintain recursive task trees with live browser progress, isolated per-conversation work items, and verification gates that require explicit acceptance evidence.MIT
- AlicenseBqualityBmaintenanceEnables any coding agent to maintain a real-time visual mindmap of its execution plan, architectural mental model, and touched files and symbols, with task statuses, directed flow links, and cross-highlighting.1428 npm1MIT
- AlicenseAqualityBmaintenanceSelf-hosted task tracker and MCP server for AI coding agents. Append-only case files preserve decisions, failed attempts, questions, and check results across sessions. A live web board lets people track progress and answer agents. Runs locally in Docker and connects to Claude Code, Codex, Cursor, and other Streamable HTTP MCP clients. MIT licensed.306MIT
- AlicenseNot gradedqualityDmaintenanceVisual architecture canvas that updates in real-time. Agents can build, read, and modify system design diagrams — services, databases, queues, APIs, entities — all linked to actual code paths in your repo.68 npm4MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.