From mistral-pack
Diagnoses and fixes Mistral AI API errors like 401 unauthorized, 429 rate limits, and invalid models using curl diagnostics, TypeScript SDK fixes, and retry backoff.
npx claudepluginhub jeremylongshore/claude-code-plugins-plus-skills --plugin mistral-packThis skill is limited to using the following tools:
Quick reference for diagnosing and fixing Mistral AI API errors. Covers HTTP status codes, SDK-specific issues, streaming failures, and tool calling problems with real solutions.
Executes Mistral AI incident triage, mitigation, and postmortem using bash scripts to check API status, rate limits, auth errors, and service health via curl and kubectl.
Guides Next.js Cache Components and Partial Prerendering (PPR): 'use cache' directives, cacheLife(), cacheTag(), revalidateTag() for caching, invalidation, static/dynamic optimization. Auto-activates on cacheComponents: true.
Guides building MCP servers enabling LLMs to interact with external services via tools. Covers best practices, TypeScript/Node (MCP SDK), Python (FastMCP).
Share bugs, ideas, or general feedback.
Quick reference for diagnosing and fixing Mistral AI API errors. Covers HTTP status codes, SDK-specific issues, streaming failures, and tool calling problems with real solutions.
MISTRAL_API_KEY configuredset -euo pipefail
# Test API connectivity and auth
curl -s -w "\nHTTP Status: %{http_code}\n" \
-H "Authorization: Bearer ${MISTRAL_API_KEY}" \
https://api.mistral.ai/v1/models | jq '.data[].id' 2>/dev/null || echo "FAILED"
# Check env
echo "Key set: ${MISTRAL_API_KEY:+yes}"
echo "Key length: ${#MISTRAL_API_KEY}"
Error: Authentication failed. Invalid API key.
Causes: Key missing, expired, revoked, or wrong workspace.
Fix:
const apiKey = process.env.MISTRAL_API_KEY;
if (!apiKey) throw new Error('MISTRAL_API_KEY is not set');
// Test the key
const client = new Mistral({ apiKey });
try {
await client.models.list();
} catch (e: any) {
if (e.status === 401) {
console.error('API key invalid — regenerate at console.mistral.ai');
}
}
Verify manually:
set -euo pipefail
curl -H "Authorization: Bearer ${MISTRAL_API_KEY}" https://api.mistral.ai/v1/models
Error: Rate limit exceeded. Retry-After: 60
Causes: Exceeded RPM (requests/min) or TPM (tokens/min) for your tier.
Fix:
async function withBackoff<T>(fn: () => Promise<T>, maxRetries = 5): Promise<T> {
for (let i = 0; i <= maxRetries; i++) {
try {
return await fn();
} catch (error: any) {
if (error.status !== 429 || i === maxRetries) throw error;
const wait = Math.min(2 ** i * 1000, 60_000);
console.warn(`Rate limited, retrying in ${wait}ms...`);
await new Promise(r => setTimeout(r, wait));
}
}
throw new Error('Max retries exceeded');
}
Check your limits: Visit console.mistral.ai/limits for workspace RPM/TPM caps.
{"message": "Unknown model: mistral-ultra"}
Fix: Use valid model IDs:
const VALID_MODELS = [
'mistral-large-latest',
'mistral-small-latest',
'codestral-latest',
'pixtral-large-latest',
'mistral-embed',
'mistral-moderation-latest',
] as const;
List available models dynamically:
set -euo pipefail
curl -H "Authorization: Bearer ${MISTRAL_API_KEY}" \
https://api.mistral.ai/v1/models | jq -r '.data[].id' | sort
{"message": "messages must be a non-empty array"}
Fix: Validate message structure before sending:
function validateMessages(messages: any[]): void {
if (!messages?.length) throw new Error('Messages array empty');
const validRoles = ['system', 'user', 'assistant', 'tool'];
for (const msg of messages) {
if (!validRoles.includes(msg.role)) {
throw new Error(`Invalid role: "${msg.role}"`);
}
if (!msg.content && !msg.toolCalls) {
throw new Error(`Message with role "${msg.role}" has no content`);
}
}
}
{"message": "tool_call_id is required for tool messages"}
Fix: Every tool result must include the matching toolCallId:
// After receiving tool_calls from the model
for (const call of response.choices[0].message.toolCalls) {
const result = await executeFunction(call.function.name, call.function.arguments);
messages.push({
role: 'tool',
name: call.function.name,
content: JSON.stringify(result),
toolCallId: call.id, // REQUIRED — must match call.id
});
}
Error: Maximum context length exceeded
Fix: Trim conversation history, keeping system message:
function trimToFit(messages: any[], maxChars = 100_000): any[] {
const system = messages.find(m => m.role === 'system');
const rest = messages.filter(m => m.role !== 'system');
const kept: any[] = system ? [system] : [];
let chars = system?.content?.length ?? 0;
// Keep most recent messages that fit
for (let i = rest.length - 1; i >= 0; i--) {
const msgChars = JSON.stringify(rest[i]).length;
if (chars + msgChars > maxChars) break;
chars += msgChars;
kept.splice(system ? 1 : 0, 0, rest[i]);
}
return kept;
}
Error: Internal server error
Causes: Mistral service issue (temporary).
Fix:
class CircuitBreaker {
private failures = 0;
private lastFailure = 0;
private readonly threshold = 5;
private readonly resetMs = 60_000;
async call<T>(fn: () => Promise<T>): Promise<T> {
if (this.failures >= this.threshold) {
if (Date.now() - this.lastFailure < this.resetMs) {
throw new Error('Circuit breaker open — Mistral service unavailable');
}
this.failures = 0; // Reset after timeout
}
try {
const result = await fn();
this.failures = 0;
return result;
} catch (error: any) {
if (error.status >= 500) {
this.failures++;
this.lastFailure = Date.now();
}
throw error;
}
}
}
Error [ERR_REQUIRE_ESM]: require() of ES Module not supported
Cause: @mistralai/mistralai is ESM-only since v1.x.
Fix: Either use import syntax (recommended) or dynamic import:
// Option 1: Convert to ESM
// package.json: "type": "module"
import { Mistral } from '@mistralai/mistralai';
// Option 2: Dynamic import in CJS
const { Mistral } = await import('@mistralai/mistralai');
Error: Request timeout after 30000ms
Fix:
const client = new Mistral({
apiKey: process.env.MISTRAL_API_KEY,
timeoutMs: 60_000, // Increase for long completions
});
// For streaming, the timeout applies to initial connection
// Individual chunks have no timeout
mistral-debug-bundle| Error | Cause | Solution |
|---|---|---|
401 | Auth failure | Regenerate key at console.mistral.ai |
429 | Rate limit | Backoff + check tier limits |
400 | Bad params | Validate model, messages, tools |
413 | Context overflow | Trim conversation history |
5xx | Service error | Retry with circuit breaker |
ERR_REQUIRE_ESM | CJS import | Use ESM import syntax |
For comprehensive debugging, see mistral-debug-bundle.