Skip to main content
Glama

Sliq

by com.getsliq
OverviewDiscussionsAdmin

Get Message Experiments

get_message_experiments
Read-only

Each test carries versions A and B with their template text, sent / replied / interested / meetings counts, and a verdict: too_early until each version has 30 sends, then the leader with its probability of scoring higher on the test's metric — leading, or wins at 95%. A send step's metric is replied. A connection request's is accepted, with an accepted count per version; there an empty version text is a request sent with no note, and recent requests haven't all been answered yet, so the version accepted faster reads a little ahead early on. Relay that probability plainly, and say "too early" when it is; at current volumes a test takes weeks. "Replied" counts any reply on that channel after the send, including replies to later steps. The tests (experiments), newest first, as {experiments: [{node_id, node_label, metric, started_at, ended_at (null while running), ended_by, versions: [{label, template_id, text, sent, replied, interested, meetings, accepted (connection requests only)}], verdict}]}, where ended_by says how a past test ended — keep_a / keep_b when that version went on alone, edit_a / edit_b when that version was edited (which restarts the test), edit otherwise — and is null while it runs; across every agent each also carries agent_id, agent_title and owner_email. The experiment at experiment_index also carries each version's messages (newest first: {prospect_id, prospect_name, person, sent_at, text, outcome}), filtered to outcome and paged 50 at a time from offset, with messages_total. A past test that sent nothing is left out, so an empty experiments means no test in that scope is running or ever sent a message.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
offsetNoWhere the page of messages starts.
node_idNoA send or connection-request step's node id, to scope to that step. Needs `agent_id`.
outcomeNoOnly messages whose prospect reached this outcome (meeting counts as interested and replied; interested counts as replied; on a connection request, replied counts as accepted). None = every sent message.
agent_idNoID of the agent (campaign) to scope to. Omit for every agent.
as_teammateNoRead a consented teammate's tests instead of your own — pass their email. Gated on that teammate's conversation-sharing setting. Omit for your own.
experiment_indexNoWhich experiment's messages to include, 0 = newest.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed3 schema fields changed
    • changedInput schema / properties / node_id / description
      Previous value: -"A send step's node id, to scope to that step. Needs `agent_id`."New value: +"A send or connection-request step's node id, to scope to that step. Needs\n`agent_id`."
    • changedInput schema / properties / outcome / anyOf
      Previous value: -[
-  {
-    "enum": [
-      "replied",
-      "interested",
-      "meeting"
-    ],
-    "type": "string"
-  },
-  {
-    "type": "null"
-  }
-]New value: +[
+  {
+    "enum": [
+      "accepted",
+      "replied",
+      "interested",
+      "meeting"
+    ],
+    "type": "string"
+  },
+  {
+    "type": "null"
+  }
+]
    • changedInput schema / properties / outcome / description
      Previous value: -"Only messages whose prospect reached this outcome (meeting counts as interested\nand replied; interested counts as replied). None = every sent message."New value: +"Only messages whose prospect reached this outcome (meeting counts as interested\nand replied; interested counts as replied; on a connection request, replied counts as\naccepted). None = every sent message."
  2. Added

TDQS

A4.6/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Goes well beyond readOnlyHint=true: explains how A/B tests run, the verdict lifecycle (too_early until 30 sends, leading, wins at 95%), per-metric definitions (replied vs accepted), the caveat that recent connection requests skew acceptance, and the weeks-long timeframe. This is unusually rich behavioral context.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads the summary, then scope, then verdict semantics. Dense but each sentence carries unique information; the returns block is long yet justified by the absence of an output schema. Minor verbosity in the metric/caveat discussion.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema and 6 parameters, the description supplies the full return shape, the ended_by vocabulary, the per-experiment messages paging, and the empty-result meaning. An agent has everything needed to call and interpret it correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so baseline is 3, but the description adds holistic meaning: how agent_id/node_id combine to set scope, that experiment_index selects which experiment's messages return, and that outcome filters messages (not the top-level list). Slightly above baseline.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource (read A/B tests of flow-step messages and connection-request notes), defines what a 'test' is operationally, and gives recognizable question phrasings. It is clearly distinct from siblings like get_campaign_flow or get_node_history.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly describes the three scoping modes (one step via agent_id+node_id, one agent, or every agent) and the questions the tool answers. It does not name a sibling alternative to route away from, so it stops short of a 5.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Sliq toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources