Claude-code-plugins supabase-common-errors

install
source · Clone the upstream repo
git clone https://github.com/jeremylongshore/claude-code-plugins-plus-skills
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/jeremylongshore/claude-code-plugins-plus-skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/plugins/saas-packs/supabase-pack/skills/supabase-common-errors" ~/.claude/skills/jeremylongshore-claude-code-plugins-supabase-common-errors && rm -rf "$T"
manifest: plugins/saas-packs/supabase-pack/skills/supabase-common-errors/SKILL.md
source content

Supabase Common Errors

Overview

Diagnostic guide for Supabase errors across PostgREST (

PGRST*
), PostgreSQL (numeric codes), Auth, Storage, and Realtime. Identify the error layer, trace the root cause, and apply the correct fix — every SDK call returns
{ data, error }
where
data
is null when
error
exists.

Prerequisites

  • @supabase/supabase-js
    installed (
    npm install @supabase/supabase-js
    )
  • SUPABASE_URL
    and
    SUPABASE_ANON_KEY
    (or
    SUPABASE_SERVICE_ROLE_KEY
    ) configured
  • Access to Supabase Dashboard (for log inspection and SQL Editor)
  • Supabase CLI installed for local development (
    npx supabase --version
    )

Instructions

Step 1 — Capture the Error Object

Every Supabase SDK call returns a

{ data, error }
tuple. Never assume
data
exists — always check
error
first.

import { createClient } from '@supabase/supabase-js'

const supabase = createClient(
  process.env.SUPABASE_URL!,
  process.env.SUPABASE_ANON_KEY!
)

// WRONG — data is null when error exists
const { data } = await supabase.from('todos').select('*')
console.log(data.length) // TypeError: Cannot read property 'length' of null

// CORRECT — always check error first
const { data, error } = await supabase.from('todos').select('*')
if (error) {
  console.error(`[${error.code}] ${error.message}`)
  console.error('Details:', error.details)
  console.error('Hint:', error.hint)
  // error.code tells you the layer:
  //   PGRST* = PostgREST (API gateway)
  //   5-digit numeric = PostgreSQL (database)
  //   AuthApiError = Auth service
  //   StorageApiError = Storage service
  return
}
// Safe to use data here
console.log(`Found ${data.length} rows`)

Troubleshooting: If

error
is undefined (not null), you may be using an older SDK version. Upgrade to
@supabase/supabase-js@2.x
or later.

Step 2 — Identify the Error Layer and Code

Match the error code prefix to the correct subsystem, then look up the specific code in the tables below.

PostgREST errors start with

PGRST
and correspond to API-layer issues (JWT, query parsing, schema). PostgreSQL errors are 5-character codes (e.g.,
42501
,
23505
) from the database engine. Auth errors come as
AuthApiError
with a human-readable message. Storage errors come as
StorageApiError
with an HTTP status.

// Diagnostic helper — paste into your codebase to classify errors automatically
function diagnoseSupabaseError(error: { code?: string; message: string; status?: number }) {
  if (!error) return 'No error'

  if (error.code?.startsWith('PGRST')) {
    return `PostgREST error ${error.code}: ${error.message}\n` +
      'Check: JWT validity, column/table names, query syntax'
  }
  if (error.code && /^\d{5}$/.test(error.code)) {
    return `PostgreSQL error ${error.code}: ${error.message}\n` +
      'Check: RLS policies, constraints, schema migrations'
  }
  if (error.message?.includes('AuthApiError')) {
    return `Auth error: ${error.message}\n` +
      'Check: credentials, email confirmation, token expiry'
  }
  if (error.message?.includes('StorageApiError')) {
    return `Storage error: ${error.message}\n` +
      'Check: bucket exists, RLS on storage.objects, file size limits'
  }
  return `Unknown error: ${JSON.stringify(error)}`
}

Troubleshooting: If the error code is empty or missing, check the HTTP status code on the response. A

401
without a code usually means
SUPABASE_ANON_KEY
is wrong or missing. A
500
without a code usually means a database function threw an unhandled exception.

Step 3 — Apply the Fix and Verify

Once you have identified the error code, apply the corresponding fix from the Error Handling table. Then verify the fix by re-running the original operation.

// Example: Fix PGRST301 (JWT expired)
// Before: stale session causes 401
const { data, error } = await supabase.from('todos').select('*')
// error.code === 'PGRST301'

// Fix: refresh the session, then retry
const { error: refreshError } = await supabase.auth.refreshSession()
if (refreshError) {
  // Token is fully invalid — force re-login
  await supabase.auth.signOut()
  console.error('Session expired. Please sign in again.')
  return
}

// Retry the original query
const { data: retryData, error: retryError } = await supabase.from('todos').select('*')
if (retryError) {
  console.error('Still failing after refresh:', retryError.code, retryError.message)
} else {
  console.log('Fixed! Retrieved', retryData.length, 'rows')
}
// Example: Fix 42501 (RLS policy violation)
// Step A: Confirm RLS is the problem using service role client
const adminClient = createClient(
  process.env.SUPABASE_URL!,
  process.env.SUPABASE_SERVICE_ROLE_KEY!,  // bypasses RLS
  { auth: { autoRefreshToken: false, persistSession: false } }
)
const { data: adminData } = await adminClient.from('todos').select('*')
console.log('Admin sees', adminData?.length, 'rows')  // If this works, RLS is blocking

// Step B: Check which user the JWT resolves to
const { data: { user } } = await supabase.auth.getUser()
console.log('Current auth.uid() =', user?.id)

// Step C: Fix the RLS policy in SQL Editor or migration
/*
  CREATE POLICY "Users can read own todos"
    ON todos FOR SELECT
    USING (auth.uid() = user_id);

  -- Verify with:
  SET request.jwt.claim.sub = '<user-id>';
  SELECT * FROM todos;
*/

// Step D: Retry original query
const { data: fixedData, error: fixedError } = await supabase.from('todos').select('*')
console.log(fixedError ? `Still blocked: ${fixedError.code}` : `Success: ${fixedData.length} rows`)

Troubleshooting: After applying a migration, you may need to reload the PostgREST schema cache. In the Supabase Dashboard, go to Settings > API and click "Reload schema cache", or call

NOTIFY pgrst, 'reload schema'
in SQL.

Output

Deliverables after applying this skill:

  • Error identified by code and layer (PostgREST, PostgreSQL, Auth, Storage, Realtime)
  • Root cause isolated using the diagnostic helper or manual code inspection
  • Fix applied from the Error Handling table and verified against the original failing operation
  • Guard code in place (
    if (error)
    checks) preventing silent null-data bugs

Error Handling

PostgREST API Errors (PGRST*)

CodeHTTPMeaningRoot CauseFix
PGRST301
401JWT expired or invalid
SUPABASE_ANON_KEY
is wrong, or the user session expired
Verify
SUPABASE_ANON_KEY
matches the project; call
supabase.auth.refreshSession()
PGRST302
401Missing Authorization headerClient created without a key, or middleware stripped the headerPass
SUPABASE_ANON_KEY
to
createClient()
; check proxy/CDN config
PGRST116
406No rows returned for
.single()
Query matched 0 rows but
.single()
expects exactly 1
Use
.maybeSingle()
for optional lookups, or check filters
PGRST200
400Invalid query parametersMalformed filter, bad operator, or invalid column referenceCheck filter syntax:
.eq('col', val)
not
.eq('col = val')
PGRST204
400Column not foundColumn name doesn't exist in the table or viewVerify column exists with
supabase gen types typescript
; check for typos
PGRST000
503Connection pool exhaustedToo many concurrent connections from serverless functionsEnable pgBouncer (Supavisor) in project settings; reduce connection count

PostgreSQL Database Errors (5-digit codes)

CodeMeaningRoot CauseFix
42501
RLS policy violationRow-level security is blocking the operation for this userAdd or fix the RLS policy; test with service role to confirm
23505
Unique constraint violationINSERT/UPDATE conflicts with an existing rowUse
.upsert({ onConflict: 'column' })
or check existence first
23503
Foreign key violationReferenced row doesn't exist in the parent tableInsert the parent row first, or check the foreign key value
42P01
Table or relation doesn't existMigration not applied, or wrong schemaRun
supabase db push
; verify schema with
\dt
in SQL Editor
42703
Column doesn't existSchema out of sync with codeRegenerate types:
supabase gen types typescript --local > types/supabase.ts
57014
Query cancelled (statement timeout)Query took longer than
statement_timeout
Add indexes; simplify the query; increase timeout in
postgresql.conf

Auth Service Errors

Error MessageCauseFix
invalid_credentials
/
Invalid login credentials
Wrong email or passwordVerify credentials; check if email is confirmed
email_not_confirmed
/
Email not confirmed
User hasn't clicked confirmation linkCheck inbox/spam; for local dev check Inbucket at
localhost:54324
user_already_exists
/
User already registered
Duplicate sign-upCall
signInWithPassword()
instead of
signUp()
Token has expired or is invalid
Stale magic link or OTPRequest a new magic link or OTP; links expire after 5 minutes by default
AuthRetryableFetchError
Network failure reaching Auth serviceRetry with backoff; verify
SUPABASE_URL
is correct and reachable

Storage Errors

ErrorCauseFix
Bucket not found
Bucket name is wrong or bucket doesn't existCreate the bucket in Dashboard or via migration SQL
The resource already exists
Uploading to a path that already has a filePass
{ upsert: true }
in upload options to overwrite
new row violates row-level security
Storage RLS blocking the upload/downloadAdd a policy on
storage.objects
for the operation (INSERT, SELECT, DELETE)
413 Payload Too Large
File exceeds the bucket's size limitIncrease
file_size_limit
on the bucket, or use TUS resumable upload for large files

Realtime Errors

SymptomCauseFix
CHANNEL_ERROR
on subscribe
Realtime not enabled for the tableDashboard > Database > Replication > enable the table; or add it to
supabase_realtime
publication
TIMED_OUT
on subscribe
Network issue or firewall blocking WebSocketCheck that port 443 WebSocket connections are allowed
No events receivedTable not in Realtime publicationRun:
ALTER PUBLICATION supabase_realtime ADD TABLE your_table;
Events stop after deploySchema change drops Realtime connectionsClients auto-reconnect; ensure
.subscribe()
handles reconnection

Examples

Example 1 — Handling
.single()
on optional data (PGRST116)

// BAD — crashes when user has no profile
const { data: profile } = await supabase
  .from('profiles')
  .select('*')
  .eq('user_id', userId)
  .single()  // throws PGRST116 if no row exists

// GOOD — returns null instead of erroring
const { data: profile, error } = await supabase
  .from('profiles')
  .select('*')
  .eq('user_id', userId)
  .maybeSingle()

if (!profile) {
  // Create a default profile
  const { data: newProfile } = await supabase
    .from('profiles')
    .insert({ user_id: userId, display_name: 'New User' })
    .select()
    .single()
}

Example 2 — Upsert to avoid unique constraint (23505)

// BAD — fails if row already exists
const { error } = await supabase
  .from('user_settings')
  .insert({ user_id: userId, theme: 'dark' })
// error.code === '23505' — unique constraint on user_id

// GOOD — inserts or updates based on conflict column
const { data, error } = await supabase
  .from('user_settings')
  .upsert(
    { user_id: userId, theme: 'dark' },
    { onConflict: 'user_id' }
  )
  .select()
  .single()

Example 3 — Realtime subscription with error handling

const channel = supabase
  .channel('todos-changes')
  .on(
    'postgres_changes',
    { event: '*', schema: 'public', table: 'todos' },
    (payload) => {
      console.log('Change received:', payload.eventType, payload.new)
    }
  )
  .subscribe((status, err) => {
    switch (status) {
      case 'SUBSCRIBED':
        console.log('Realtime connected')
        break
      case 'CHANNEL_ERROR':
        console.error('Realtime error — is the table in the publication?', err)
        // Fix: ALTER PUBLICATION supabase_realtime ADD TABLE todos;
        break
      case 'TIMED_OUT':
        console.error('Realtime timed out — check network')
        break
      case 'CLOSED':
        console.log('Channel closed')
        break
    }
  })

// Always clean up on unmount / exit
process.on('SIGINT', async () => {
  await supabase.removeChannel(channel)
  process.exit(0)
})

Example 4 — Connection pool exhaustion (PGRST000) in serverless

// BAD — creates a new client per request in serverless (Lambda, Edge Functions)
export async function handler(req: Request) {
  const supabase = createClient(url, key)  // new connection every invocation
  const { data } = await supabase.from('todos').select('*')
  return Response.json(data)
}

// GOOD — reuse client across warm invocations
const supabase = createClient(url, key, {
  auth: { autoRefreshToken: false, persistSession: false }
})

export async function handler(req: Request) {
  const { data, error } = await supabase.from('todos').select('*')
  if (error) {
    if (error.code === 'PGRST000') {
      // Pool exhausted — return 503 so the caller retries
      return new Response('Service temporarily unavailable', { status: 503 })
    }
    return Response.json({ error: error.message }, { status: 400 })
  }
  return Response.json(data)
}

Resources

Next Steps

  • Use
    supabase-debug-bundle
    to generate a full diagnostic snapshot when errors persist after applying these fixes.
  • Use
    supabase-security-basics
    to audit your RLS policies and prevent
    42501
    errors proactively.
  • Use
    supabase-known-pitfalls
    for edge cases and SDK behavior that can cause subtle bugs.
  • Use
    supabase-observability
    to set up logging and alerting so you catch errors before users report them.