Skip to main content
Glama
Sign In

autario-mcp

by io.github.Autario
OverviewDiscussionsAdmin

outreach_prospects

ADMIN/CURATOR ONLY, the outreach engine's prospects (people who are NOT accounts, mailed a paced 3-step sequence from hello@mail.autario.com; spec 2026-11-08-outreach-engine). The prospects are the agencies and merchants of the PMF test (the people who run Shopify, Meta Ads and GA4 for clients), never autario accounts. Four actions: add inserts rows for a campaign key ({email, country, segment, source_url, name?, company?, variant?}, max 500 per call): addresses are lower-cased, the domain derived, the claim variant A/B assigned round-robin per segment when missing, utm fields stamped (utm_source=outreach, utm_medium=email, utm_campaign=, utm_content=), and every row that fails eligibility right now is reported back with its reason (country_not_allowed outside US/GB/CA/AU by default, opted_out, domain_suppressed, is_account, duplicate) instead of sitting in the table; list returns prospects filtered by campaign and status; stats returns per campaign and variant the counts by status, sends per step, opened, clicked and signups (prospects whose address became an account), today's sends, the daily cap and ramp day, and whether the sending domain is verified; mark sets one prospect to replied or blocked by hand (reply detection is manual); history returns ONE prospect (by prospect_id or email) with every mail sent to it, the provider events per mail (delivered, opened, clicked with the link, bounced, complained), first delivered/opened/clicked per mail and the signup date when the address became an account. list also filters by segment, country, variant and a free-text q, and every row carries delivered_at, opened_at, clicked_at, clicks and signed_up_at. start releases the next batch of a campaign: batch_size waiting prospects per variant (default 300), and it needs confirm_count = the planned release (a call without it, or with a stale number, answers the planned count per variant and releases nothing), the same door as the Start button on the Sequence tab; a release only gives mail 1 a due date, the worker sends it paced under the daily cap and only while OUTREACH_MAIL is on (refused while the switch is off, without a postal address or with an unverified sending domain); calling it again on a running campaign releases the next batch. The same doors as /api/admin/outreach. Requires the connector to be OAuth-authorized as the autario curator account | any other caller gets a permission error. Use when asked "add these prospects", "how is the outreach going", "mark this one as replied", "release the next 300".

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "action": "<action>"
    },
    "name": "outreach_prospects"
  }
}

Input Schema

NameRequiredDescriptionDefault
qNolist only: free text matched against email, company, name and domain.
rowsNoadd only: the prospects to insert. Each {email, country (ISO alpha-2), segment (agency | ecommerce | ai_user | free text), source_url (where the address is published), name?, company?, variant? (A | B)}.
emailNohistory only: the prospect address, instead of prospect_id.
limitNolist only: max rows (default 100).
actionYesadd = insert prospects for a campaign; list = prospects by campaign/status/segment/country/variant/q; stats = per campaign and variant counts (waiting, released), opened, clicked, signups, cap, ramp day, domain state; mark = set one prospect by hand; history = one prospect with its mails, events and signup date; start = release the next batch of a campaign (batch_size per variant, confirm_count required).
formatNoOutput wire format for this MCP call. Default 'toon' (Token-Oriented Notation, fewest tokens, best for tabular rows). 'compact' = minified JSON. 'json' = pretty JSON for readability. The REST API always returns JSON regardless.
statusNolist: filter by a prospect status word; mark: the status to set, replied or blocked only.
countryNolist only: filter by ISO alpha-2 country.
segmentNolist only: filter by segment (exact).
variantNolist only: filter by claim variant A or B.
campaignNoThe sequence key the drafts carry (e.g. outreach-pmf-2026-11). Required for add and start, optional filter for list.
batch_sizeNostart only: waiting prospects to release per variant (default 300).
prospect_idNomark and history: the prospect id from list.
confirm_countNostart only: the total this batch releases, as the door answered it; a missing or stale number releases nothing and answers the planned count.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.7/5.0
Behavior5/5

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

Goes well past the annotations, which only establish non-read-only, non-idempotent, non-destructive, non-open-world. The description discloses normalization side effects (lower-casing, domain derivation), round-robin variant assignment, utm stamping, per-row eligibility rejections with named reasons (country_not_allowed, opted_out, domain_suppressed, is_account, duplicate), the confirm_count stale-number guard that releases nothing, paced sending under a daily cap, and refusals when OUTREACH_MAIL is off, no postal address exists, or the sending domain is unverified. It also flags that reply detection is manual, which is exactly the kind of caveat an agent would otherwise get wrong.

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

Conciseness3/5

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

Front-loaded with the admin restriction and resource definition, and nearly every clause carries actionable information, but the whole thing is a single semicolon-chained block that is hard to scan for a six-action, fourteen-parameter tool. The 'Four actions:' label that precedes five names, plus the late-arriving start paragraph, cost it structure points rather than content points.

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?

There is no output schema, so the description carries the return-value burden itself — and it does, describing what list rows carry (delivered_at, opened_at, clicked_at, clicks, signed_up_at), what stats aggregates (per-campaign/variant status counts, sends per step, opened, clicked, signups, today's sends, cap, ramp day, domain verified), and what history returns per mail (provider events, first delivered/opened/clicked, signup date). Combined with auth requirements and failure modes, an agent has everything needed to call 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 description coverage is 100%, so the baseline is 3, but the description earns more by explaining semantics the schema only gestures at: confirm_count must match the planned release or nothing is sent and the planned count is returned, batch_size is per variant, campaign is required for add/start but only a filter for list, and mark accepts only replied/blocked. It adds a little less on rows fields (variant, country) than it could, since those are already fully documented in the schema.

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?

Names a specific resource (the outreach engine's prospects, explicitly 'people who are NOT accounts') and enumerates each action's effect: add inserts rows, list returns filtered prospects, stats returns per-campaign/variant counts, mark sets replied/blocked, history returns one prospect's mail trail, start releases a batch. Nothing in the sibling list overlaps, so an agent can route here without ambiguity — though the phrase 'Four actions:' followed by five action names is a miscount that slightly muddies the enumeration.

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

Usage Guidelines5/5

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

Gives explicit trigger phrases ('add these prospects', 'how is the outreach going', 'mark this one as replied', 'release the next 300'), states per-action preconditions (campaign required for add/start, confirm_count required for start, prospect_id or email for history), and names the exclusion up front: ADMIN/CURATOR ONLY, requiring OAuth as the autario curator account, with any other caller receiving a permission error.

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

All autario-mcp 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