What did Hlido score crimson-crab?

crimson-crab scored 76/100 (STEADY) in Hlido's independent, hands-on review.

How much does crimson-crab cost?

Hlido recorded the pricing model as: Open source. Derived from evidence Hlido holds, not vendor-supplied; verify current prices on the vendor's pricing page.

Does any vendor pay Hlido for placement?

No. Hlido takes no money from the agents it rates — scoring weights stay private and the evidence behind every verdict is public.

crimson-crab

Name: crimson-crab review
Item: crimson-crab
Rating: 76
Author: Hlido Editor

Frameworks & Eval · tested 2026-08-09 · re-test due 2026-11-09 · by the Hlido desk, not the vendor

In short: A Rust Claude SDK built on one compile-time guarantee — no unwrap, expect, panic! or todo! anywhere in the library — with unknown API surface deserialising into catch-alls instead of breaking builds.

6 PASS · 1 FAIL of 7 public-surface claims

Visit crimson-crab → scorecard.json

Quick answer

crimson-crab scores 76/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-08-09). STEADY (76) for a genuinely checkable core guarantee (panicking macros denied at compile time, so surprises surface as errors), forward-compatible catch-all deserialisation that survives additive API changes, and complet Pricing: Open source (free entry point documented).

Hlido Editor · 2026-08-09

crimson-crab makes an unusually specific and checkable promise for an SDK: panicking macros are denied at compile time across the whole library, so a malformed or surprising API response reaches the caller as an `Error` rather than taking down a service. That is a lint-enforceable property rather than a claim about care, which is what makes it worth something. The second design decision is forward compatibility — unknown block types, stream events and stop reasons deserialise into `Unknown` catch-alls, so a new model or content type from Anthropic does not break a build on the day it ships. Anyone who has had a strongly-typed client break on an additive API change will recognise the value. Coverage is the complete Claude surface rather than a subset: streaming via a hand-rolled SSE parser with no eventsource dependency, tool use with raw JSON schemas, adaptive/budgeted/disabled thinking modes, prompt caching with 5m/1h TTL and cache hits readable off `Usage`, message batches with a JSONL-decoding result stream, token counting, and typed errors carrying `request_id` with jittered backoff honouring retry-after. The public API exposes `futures_core::Stream` rather than tokio types, which keeps it runtime-light and checks on wasm32. The caution is maturity and standing: v0.3.0 is early, MSRV 1.75, and this is a community SDK rather than an Anthropic-published one — the surface carries no adoption numbers, no named users and no third-party review. The slug also describes it as an 'MCP template', which the captured surface does not support: this is a Claude API client.

Why STEADY

STEADY (76) for a genuinely checkable core guarantee (panicking macros denied at compile time, so surprises surface as errors), forward-compatible catch-all deserialisation that survives additive API changes, and complete rather than partial Claude API coverage including streaming, batches, caching and token counting. Held well below VITAL by v0.3.0 maturity, no adoption or third-party validation on the surface, and unofficial status — a community SDK carries different risk from a vendor-published one.

Public-surface checklist

PASS Homepage loads (required)
PASS Primary value prop (required) — 'The Rust SDK for Claude that can't panic on you.'
PASS Cta present (required) — `cargo add crimson-crab` with copy button; crates.io and docs.rs links
PASS Pricing or access — Open source, MIT OR Apache-2.0
PASS Docs present (required) — docs.rs plus end-to-end examples for basic, streaming, tool runner and typed output
PASS Version disclosed — v0.3.0, MSRV 1.75 stated on the page
FAIL Third party validation — No adoption figures or independent review

What we saw

4 screenshots captured by the Hlido engine during the reviewed run (run-cd0dae69dc2f1ff0-singhpratech-github-io). Our own captures — not vendor marketing material.

crimson-crab — run screenshot 1 (home.png) — home.png

crimson-crab — run screenshot 2 (page_main.png) — page_main.png

crimson-crab — run screenshot 3 (page_top.png) — page_top.png

crimson-crab — run screenshot 4 (page_features.png) — page_features.png

What it does well

unwrap, expect, panic! and todo! denied at compile time across the library — a lint-enforced guarantee, not a promise
Unknown block types, stream events and stop reasons deserialise into catch-alls, so additive API changes do not break builds
Complete Claude surface: streaming, tool use, thinking modes, prompt caching, batches, token counting
Hand-rolled SSE parser yields fine-grained events and accumulates a final message, with no eventsource dependency
Typed errors carry request_id; backoff with jitter honours retry-after automatically
Public API exposes futures_core::Stream rather than tokio types — runtime-light and checks on wasm32
Prompt caching exposes 5m/1h TTL and reads cache hits back off Usage
Permissive dual licence (MIT OR Apache-2.0) with a stated MSRV of 1.75

What it fails at

v0.3.0 — early, with no changelog or stability commitment on the surface
Community SDK, not Anthropic-published; no third-party review or named users
No adoption numbers, download figures or production references
The registry slug calls it an 'MCP template', which the captured surface does not support — it is a Claude API client
Not tested hands-on by Hlido; the no-panic guarantee is a build-time property that deserves verification

Best for

Rust services calling Claude where a panic in a dependency is an availability incident
Teams that need the full Claude surface (batches, caching, thinking, streaming) rather than a chat-completions subset
WASM or non-tokio runtimes that a tokio-typed client would not fit

Not recommended for

Anyone requiring an officially supported vendor SDK
Projects that cannot take a pre-1.0 dependency on a critical path
Anyone expecting an MCP server or template — this is a Claude API client

Pricing & access

ModelOpen source
Free entry pointYes — a free tier or open-source edition is documented
Pricing findable on the public surfacePASS Open source, MIT OR Apache-2.0 (tested 2026-08-09)

Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-08-09.

Related agents

Agent relevance

API SDK Behavioral-testable

Agentic-Commerce Readiness 62/100 · INTEGRABLE

Independent readiness for agent delegation & transaction. How it’s scored · check live

A Rust SDK for building agents rather than an agent itself: tool definitions take raw JSON schemas, the agentic loop is driven by the caller, and any server-tool passes through ToolUnion::Raw. Relevant to agent builders on Rust; nothing here is agent-callable at runtime.

Agent-friendly score: 6/10

Evidence

unwrap, expect, panic! and todo! denied at compile time across the library — source (2026-08-09) verified
Unknown block types, stream events and stop reasons deserialise into Unknown catch-alls — source (2026-08-09) verified
Covers streaming, tool use, adaptive thinking, prompt caching, batches and token counting — source (2026-08-09) verified
Public API exposes futures_core::Stream, not tokio types; checks on wasm32-unknown-unknown — source (2026-08-09) verified
v0.3.0, MIT OR Apache-2.0, MSRV 1.75, published on crates.io — source (2026-08-09) verified
Officially supported by Anthropic — source (2026-08-09)
Adoption numbers, named users or third-party review — source (2026-08-09)

scorecard.json · transparency passport · registry · methodology

More: compare agents · best of · developer tools · incident registry

Verdict by Hlido Editor, our automated editorial system · Method: public-surface-tier-2+editorial-narrative-v2 · Methodology version 2026.05 · Next review due 2026-11-09

How this page was produced. The scores, claim verdicts and evidence come from automated hands-on testing of the product’s public surface. The written analysis is drafted by an AI system, and pages publish without a person reviewing each one. Hlido publishes this record and answers for it — tell us if anything here is wrong and we will correct it.

Embed this trust badge

Live, always-current independent score — free to embed on your site or README. No vendor pays for placement.

Markdown

[![Hlido trust score](https://hlido.eu/badge/singhpratech-crimson-crab-mcp-template.svg)](https://hlido.eu/check/?agent=singhpratech-crimson-crab-mcp-template)

HTML

<a href="https://hlido.eu/check/?agent=singhpratech-crimson-crab-mcp-template"><img src="https://hlido.eu/badge/singhpratech-crimson-crab-mcp-template.svg" alt="Hlido trust score"></a>