Boosthis

Session Summary

boosthis_session_summary

Read-only

Per-screen p50/p75/p95 and worst rating, worst screens first (with p99, spike ratio and stdev spread where the server has them). It answers live from the project's own Boosthis server when read credentials are configured (BOOSTHIS_INSTALL_ID + BOOSTHIS_READ_TOKEN from the in-app dashboard, full-details telemetry on), and on the hosted server from install_id + read_token arguments. Without them the answer is a note pointing at the in-app dashboard, not an empty result. An account can hold several projects, each reporting in several runtimes. With an account_token, project (name, 'name (runtime)', or an install id) selects one; without it the answer is the most recent reporter. The rest are listed under your_projects; entries sharing a project_group are one project in different runtimes, each with its own readings.

Input Schema

TableJSON Schema

Name	Required	Description
`project`	No	Which project in the account to read — its dashboard name, that name with a runtime as 'name (runtime)', or an install id. Only meaningful with account_token.
`install_id`	No	Optional: the install id to read live data for. On the HOSTED Boosthis MCP, copy it from the in-app dashboard's "Connect your AI" card and pass it here. Omit on a local stdio server (it uses BOOSTHIS_INSTALL_ID from the env).
`read_token`	No	Optional: the SELF-scoped read token for that install (paired with install_id). It is read-only — it can read this app's own perf data but CANNOT delete it. Copy it from the in-app dashboard. Omit on a local stdio server (it uses BOOSTHIS_READ_TOKEN from the env).

TDQS

A4.6/5.0

Behavior5/5

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

Beyond the readOnlyHint annotation, it reveals the live-vs-note behavior, that credentials gate real data, that an account can hold multiple projects, and how project selection behaves with and without an account_token. It also discloses that read_token is self-scoped and cannot delete data, which materially informs an agent's expectations.

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 a dense single paragraph, but it is front-loaded with the core result format and every subsequent sentence conveys a distinct behavior (credentials, fallback, multi-project, project grouping). It is longer than ideal, yet the complexity of the tool justifies the length.

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, the description carries the burden of explaining return values, and it does: per-screen quantiles, worst-first ordering, and the `your_projects` listing. It also covers all three optional parameters, the two server modes, and the no-credentials fallback, leaving no significant gap for an agent 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?

The input schema already covers all three parameters, so the baseline is 3. The description adds value by specifying when each parameter should be omitted (local stdio vs hosted MCP) and how `project` relates to account_token and runtime-qualified names, which is not fully captured 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?

The first sentence names a specific resource ('Per-screen p50/p75/p95 and worst rating') and a clear action ('answers live'), so an agent immediately knows what the tool returns. The 'worst screens first' ordering and the optional p99/spike/stdev extensions make it distinguishable from sibling snapshot/trend tools without opening 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 Guidelines4/5

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

The description explains exactly when live data is available (configured read credentials vs. install_id/read_token arguments), and states the fallback behavior ('a note pointing at the in-app dashboard'). It does not explicitly name alternative sibling tools or say 'use X instead', but the context is clear and actionable.

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

All Boosthis tools Inspect

Try in Browser

RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

TDQS

A3.9/5.0

Disambiguation4/5

Most tools target a distinct concern—alerts, budgets, jobs, promises, rules, crash risk, install/removal—and the descriptions are detailed enough to separate them. A few pairs, like verify_kit/verify_kit_install and recent_samples/session_summary, could be misselected without close reading, but they are not functionally identical.

Naming Consistency3/5

All tools share the boosthis_ prefix and use snake_case, which creates a recognizable namespace. However, the second part mixes noun-only names (alerts, budgets, jobs, snapshot, trend), verb-based names (get_rule, remember_promise, verify_kit), and one sentence-style name (what_should_i_look_at_next), so there is no consistent verb_noun convention.

Tool Count3/5

At 24 tools, this sits squarely in the borderline-heavy range for an MCP server. Each tool does have a distinct role, but the surface feels large, especially with several overlapping read-only diagnostics that could potentially be consolidated.

Completeness4/5

The set covers the main observability lifecycle well: install, verify, monitor, diagnose, check trends, and record promises. Minor gaps exist—like no way to mutate alert states or delete promises through the MCP—but these appear intentional and are documented as dashboard-side actions.

Resources

Need Help?