Skip to main content
Glama

configure_ad_step

Configure ONE step of an ad's setup wizard. CRITICAL: configuring the SUMMARY step FINISHES + ACTIVATES the ad — in AUTO product mode that publishes EVERY matching product to the LIVE channel (spends budget), so to test safely set MANUAL mode FIRST (set_ad_product_mode) and submit one item with run_ad_item_action; check activationBehavior in get_ad before finishing. Ads are ads-based channels (advertising / price-comparison / classifieds), not marketplaces, and have no order sync. Ads are set up step by step IN ORDER: call this repeatedly, targeting each step the previous result reported as nextStepType, until nextStepType is null (setup complete). Identify the step by step_index (from get_ad steps[].index) — REQUIRED to disambiguate when two steps share a step_type (e.g. two 'form' steps); step_type alone targets the FIRST step of that type. Configuring out of order or before prerequisites are met returns 'prerequisite_not_met' / 'step_not_available' (e.g. category mapping precedes the attribute step). PAYLOAD: for a form / order-settings step, payload = {items: {fieldCode: value, ...}} built from that step's fields (get_ad steps[].fields — each field is {code, label, type, required, options, optionsSource, htmlAttributes, value, help}); send ONLY the fields you want to change — the others keep their current value and nothing is deleted. A dropdown/options field takes the option's VALUE — options is a list of {value, label} rows, so send options[].value, NEVER a row's position in that list; If a field arrives with optionsPending=true its option list is not known yet (usually because the channel connection does not exist); finish the auth step and re-read get_ad rather than guessing a value. A REQUIRED field with an empty value is yours to fill even when htmlAttributes look pre-filled — the server only supplies it when it can resolve it from the connection, and it says so in the error when it cannot. read steps[].fields[].options first. Structured steps use their own payload: the product-filter step accepts attribute_filter (the condition tree as a JSON string, or "" to clear it), category_filter (a JSON string {categories_operator, categories:[...]}), and the flag export_out_of_stock ("0"|"1"); send ONLY what changes — it merges over the current filter, so the untouched flags/filters are kept. To include or exclude out-of-stock products use the export_out_of_stock flag — NEVER put a qty/stock condition in attribute_filter (it is an invalid state and is rejected). (export_children_as_parent and export_removed_products are managed in the Koongo UI and CANNOT be set here — sending either is rejected.) category-mapping and mapping-table likewise take their own payload. Returns {integrationId, stepType, status, ready, lastStep, nextStepType, oauthUrl, transitions[], message}. When transitions[] reports a product export/generation (e.g. moving into the summary/preview), that runs ASYNC — poll ad_status until productsRefreshing AND productsSubmitting are false before reading get_ad_report (its dataCheckErrors = the KOONGO DATA CHECK) or trusting counts. status: 'saved' (applied; transitions[] lists automation that ran), 'oauth_required' or 'credentials_in_ui' (a CONNECTION step — see below; TERMINAL, do not retry), 'prerequisite_not_met' / 'step_not_available', or 'validation_failed' (the message lists the field problems — unknown field, value outside its options, non-numeric value, or missing required field — fix them using the step's fields). CATEGORY ATTRIBUTES MAY BE DEFERRED: on channels whose attributes depend on the mapped categories, saving the product-filter step starts that download in the BACKGROUND instead of doing it inline. You can tell: transitions[] then says the download was started (and does NOT claim the attributes were fetched or an export ran), and categoryAttributeWarmup is present with pending:true. Do not call the refresh tool and do not read the attribute list yet — poll ad_status until the block is absent or reports state 'ready'. The attributes are pulled in automatically when it finishes; state 'failed' means call refresh_ad_category_attributes once. CONNECTION IS NOT CONFIGURABLE HERE: the connection/auth step (credentials, API keys, OAuth, auth settings) can NEVER be set through MCP — this is intentional and permanent for security, not a limitation to work around. When a step returns 'oauth_required' or 'credentials_in_ui', that status is FINAL for that step: do NOT resend it, do NOT guess field codes or a payload, do NOT keep retrying — the result will not change (it is not transient and not a validation error you can fix). Hand the returned oauthUrl (OAuth) or editUrl (credentials, entered in the Koongo UI) to the USER, then STOP and wait; once they finish, re-check ad_status (isConnected) and continue with the NEXT step. ad_id is from list_ads / create_ad. project_id is OPTIONAL (inferred for a single-project customer; project_id_required otherwise — then call list_projects).

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
ad_idYes
payloadNoSettings for this step. Form / order-settings: {items: {fieldCode: value}} from get_ad steps[].fields. A connection/auth step takes NO payload and cannot be configured here — omit it (also for advance-only steps).
step_typeNoTarget the step by type (the nextStepType from the previous step). Ambiguous when two steps share a type — prefer step_index.
project_idNo
step_indexNoTarget the step by its index (get_ad steps[].index) — use this to disambiguate steps that share a step_type.

Schema Changelog

Changes observed during successful MCP inspections. Dates show when Glama detected each change.

  1. First observed

TDQS

A4.9/5.0
Behavior5/5

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

Annotations are minimal (readOnlyHint false, destructiveHint false), so the description carries the burden — and it over-delivers. It discloses the budget-spending activation side effect of the SUMMARY step, async product export/generation in transitions[], background category-attribute downloads, terminal (non-retryable) oauth/credentials statuses, and the reject-list of unsettable flags. It also corrects a plausible misreading of destructiveHint=false by warning that AUTO mode publishes to the LIVE channel and spends budget. No contradiction with annotations, which simply understate this risk.

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?

The description is very long, but nearly every sentence carries unique operational information for a genuinely complex tool with heterogeneous step types, async transitions, and deferred attribute loading. It is front-loaded with the critical activation warning and ordering rule before payload details. Minor redundancy exists — the non-configurability of the connection step and the do-not-retry instruction are stated two or three times — which costs it full marks but is arguably intentional emphasis.

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 correctly documents the return shape ({integrationId, stepType, status, ready, lastStep, nextStepType, oauthUrl, transitions[], message}) and explains every status value, the async polling requirements, terminal vs retryable conditions, and the success path. For a tool this complex, nothing an agent needs to call it safely and correctly is missing.

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

Parameters5/5

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

Schema coverage is only 60%, and the description compensates massively. It explains ad_id provenance (list_ads/create_ad), when project_id is required and how to resolve it (list_projects), the step_index vs step_type disambiguation rule, the exact payload shape {items: {fieldCode: value}} with send-only-changed semantics, dropdown value-vs-position rule, optionsPending handling, required-field behavior, and dedicated payloads for the product-filter step (attribute_filter, category_filter, export_out_of_stock) including what gets rejected.

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?

Opens with a specific verb + resource + scope: "Configure ONE step of an ad's setup wizard." The emphasis on "ONE step" and the repeated contrast with ads-based channels ("not marketplaces, and have no order sync") clearly distinguishes it from siblings like configure_marketplace_step, set_ad_product_mode, and set_ad_name without needing to open their schemas.

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?

Exceptionally explicit about when and when not to use the tool: it names set_ad_product_mode for safe testing, says to poll ad_status instead of calling the refresh tool while attributes warm up, routes state 'failed' to refresh_ad_category_attributes once, and states the connection/auth step "can NEVER be set through MCP." It even prescribes the stop-and-wait user handoff for oauthUrl/editUrl. The description also documents the step-by-step iteration loop terminating at nextStepType null.

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

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

TDQS

A4.2/5.0
Disambiguation4/5

Tools are organized around distinct resources (ads, marketplaces, feeds, orders, rules, sources) with clear action verbs, and descriptions explicitly disambiguate near-pairs like get_feed_status vs ad_status or set_feed_filter vs set_feed_attribute_filter. A few similarly named status/action pairs (e.g. ad_status vs get_ad, run_ad_item_action vs run_ad_operation) require careful reading, but overall the purposes are separable.

Naming Consistency4/5

The overwhelming majority follow a consistent verb_noun snake_case pattern (list_*, get_*, create_*, set_*, run_*, test_*). Minor deviations like ad_status and marketplace_status (noun-based status tools) and koongo_knowledge break the pattern slightly, but the convention is clearly recognizable and predictable.

Tool Count1/5

At 105 tools, the surface is extreme and far beyond the 50+ threshold, even for a complex e-commerce integration domain. Much of the bloat comes from systematic triplication across ads, marketplaces, and feeds (e.g. three nearly identical map_*_attribute tools, three list_*_items, three get_*_report) that a generic resource parameter could have consolidated.

Completeness4/5

The toolset covers the full lifecycle of feeds, ads, marketplaces, order connections, rules, and imports, including create/read/update/delete, status monitoring, item-level actions, validation, repair, and restore. Minor gaps exist, such as no delete for standalone order connections and limited update capabilities for some entities, but these are workable and do not create dead ends for the core workflows.

Resources