MCP tool catalog

The Metriq MCP server exposes 44 tools (14 read, 30 write). Connect any MCP-compatible agent (Claude Desktop, Claude Code, or your own runtime) and it can read your board's flow analytics and drive work items. Each tool is team-scoped: the team is derived from the API key, so a call can never address another team's data.

This page is generated from the server source. For setup, the auth model, workflow recipes, and the agent end-of-turn contract, see the MCP reference.

Read tools

Read tools never mutate the board and never broadcast updates.

ToolSummary
get_bless_telemetryBless-latency (p50/p90/max in ms) and acceptance-rate (blessed / decided) for agent-Done blessing requests in a window.
get_blockersCurrently blocked work items (with how long each has been blocked) plus the per-week count of new blocked events over the last 4 weeks.
get_boardGet the full board state with all columns (Backlog, Ready, In Progress, Done) and their work items.
get_cost_leverage_telemetryInternal/admin instrumentation read (no user-facing billing).
get_dispatch_journeyReconstruct the full ordered lifecycle of a SINGLE dispatch by its dispatch_id (the same id wait_for_assignment returns and claim_assigned_task takes).
get_dispatch_telemetryDispatch funnel telemetry over a window: time_to_first_claim p50/p95 (delivered_at - created_at), time_to_first_artifact p50/p95 (artifact.posted_at - claimed_at), dispatch_to_blessing_request p50/p95 (review_requested - created_at), plus counters dispatched_total / claimed_total / artifact_reported_total / redelivered_total / stuck_total / superseded_total.
get_flow_summaryHeadline flow-health digest: WIP per column, throughput trend, cycle time percentiles, SLE state, blockers, stale items.
get_forecastMonte Carlo forecast based on the team's throughput history.
get_itemGet detailed information about a single work item, including its full event history.
get_pending_inputRead the human↔agent conversation thread for a work item the agent asked about.
list_active_dispatchesEnumerate dispatches that are currently in flight or were recently completed for the actor's team.
list_labelsList all labels for the team.
list_membersList all members of the team.
request_human_inputExplicitly request human input on a work item the agent is working on.

Write tools

Write tools mutate the board; on success they broadcast a live update to any open Metriq tab.

ToolSummary
add_checklist_itemAdd a checklist item under a work item.
archive_itemArchive a work item.
assign_itemSet the assignees for a work item.
block_itemToggle the blocked status of a work item.
claim_assigned_taskAtomically claim a delivered dispatch for a specific agent_session_id.
clear_item_outcomeRemove the outcome a work item serves (if any).
clear_label_outcomeDemote an outcome label back to a regular label, clearing its problem statement.
create_itemCreate a new work item in the Backlog column.
create_labelCreate a new label scoped to the actor's team.
delete_checklist_itemRemove a checklist item from a work item.
delete_itemPermanently delete a work item.
demote_epicDemote an epic back to a regular label.
edit_checklist_itemChange the text of a checklist item.
end_agent_sessionGraceful shutdown of an agent session: sets agent_sessions.ended_at = now().
move_itemMove a work item to a different column.
post_agent_commentPost an agent self-narration comment on a work item.
promote_label_to_epicPromote a label to epic status.
register_agent_sessionRegister or rejoin an agent session for the actor's team.
report_artifactReport a URL pointing at a reviewable artifact (PR, branch, commit, diff, Loom, gist — anything that returns a URL).
report_token_usageReport the LLM token usage for the most recent agent round-trip.
set_item_labelsReplace the label set on a work item with exactly these labelIds.
set_item_outcomeSet the single outcome a work item serves, REPLACING any outcome already on the card (a card serves at most one outcome).
set_label_as_outcomePromote a label into the 'outcome' kind and set its one-line problem/opportunity statement.
set_learning_outcomeRecord whether a shipped (Done) card moved the outcome it served: validated | invalidated | inconclusive | unset.
set_outcome_metricSet the ONE metric on an outcome label: a metricName plus a baseline and a current numeric value, all entered manually (there is no automatic ingestion).
toggle_checklist_itemFlip the checked state of a checklist item.
unarchive_itemRestore an archived work item to its previous state.
update_itemEdit a work item's title, description, type, or priority.
update_outcome_currentUpdate ONLY the current value of an outcome's metric (the baseline is fixed).
wait_for_assignmentLong-poll for the next pending dispatch assigned to this team's agent fleet.