Neo4j Driver Go Skill

neo4j-contrib/neo4j-driver-go-skill

Covers the Neo4j Go Driver v6 — driver lifecycle, ExecuteQuery, managed and explicit transactions, session config, error handling, data type mapping, and connection tuning. Use when writing Go code that connects to Neo4j, setting up NewDriver or ExecuteQuery, debugging sessions/transactions/result handling, or working with neo4j-go-driver v5→v6 migration. Triggers on NewDriver, ExecuteQuery, SessionConfig, ManagedTransaction, neo4j-go-driver. Does NOT handle Cypher query authoring — use neo4j-cypher-skill. Does NOT cover driver version migration steps — use neo4j-migration-skill.

5k tokens

context cost

the whole folder, loaded on every use

files

instructions only

copies elsewhere

how many repositories repackaged it

101

stars on the repo

on the repository, not the skill itself

Install

one command, takes just this skill from the repository

npx skills add https://github.com/neo4j-contrib/neo4j-skills --skill neo4j-driver-go-skill

What comes with it

4 621 bytes besides the instruction

README.md

references/advanced-config.md

references/repository-pattern.md

What it tells the agent to use

found in the instruction text

Bash runs shell commands — read the instruction before connecting

WebFetch fetches pages from the network

The instruction itself

22 sections, as written by the author

When to Use

Writing Go code that connects to Neo4j
Setting up neo4j.NewDriver(), ExecuteQuery(), or session/transaction patterns
Debugging connection errors, result iteration, type assertions, causal consistency

When NOT to Use

Writing/optimizing Cypher → neo4j-cypher-skill
v5→v6 migration steps → neo4j-migration-skill

Installation

go get github.com/neo4j/neo4j-go-driver/v6

Import: github.com/neo4j/neo4j-go-driver/v6/neo4j

v5→v6 rename (deprecated aliases still compile, remove before v7):

| v5 | v6 |

|----|----|

| neo4j.NewDriverWithContext(...) | neo4j.NewDriver(...) |

| neo4j.DriverWithContext | neo4j.Driver |

Environment Variables

import "os"

uri      := getEnv("NEO4J_URI",      "neo4j://localhost:7687")
user     := getEnv("NEO4J_USERNAME", "neo4j")
password := getEnv("NEO4J_PASSWORD", "")
database := getEnv("NEO4J_DATABASE", "neo4j")

func getEnv(key, fallback string) string {
    if v := os.Getenv(key); v != "" { return v }
    return fallback
}

Use godotenv to load .env in dev: godotenv.Load(). .env in .gitignore.

Driver Lifecycle

One Driver per application. Goroutine-safe, connection-pooled, expensive to create.

func NewNeo4jDriver(uri, user, password string) (neo4j.Driver, error) {
    driver, err := neo4j.NewDriver(
        uri, // "neo4j+s://xxx.databases.neo4j.io" for Aura
        neo4j.BasicAuth(user, password, ""),
    )
    if err != nil {
        return nil, fmt.Errorf("create driver: %w", err)
    }
    ctx := context.Background()
    if err := driver.VerifyConnectivity(ctx); err != nil {
        driver.Close(ctx)
        return nil, fmt.Errorf("verify connectivity: %w", err)
    }
    return driver, nil
}

// In main / app teardown:
defer driver.Close(ctx)

❌ Never create driver per-request. Create once at startup; share across goroutines.

URI schemes: neo4j+s:// (Aura/TLS+routing), neo4j:// (plain+routing), bolt+s:// (TLS+single), bolt:// (plain+single).

Choosing the Right API

|-----|----------|:----------:|:------------:|

| session.ExecuteRead/Write() | Large result sets / streaming | ✅ | ✅ |

| session.BeginTransaction() | Spans multiple functions / ext coordination | ❌ | ✅ |

| session.Run() | CALL IN TRANSACTIONS / auto-commit only | ❌ | ✅ |

CALL { … } IN TRANSACTIONS and USING PERIODIC COMMIT manage their own transactions — use session.Run(). They fail inside managed transactions.

ExecuteQuery (Recommended Default)

Manages sessions, transactions, retries, and bookmarks automatically.

result, err := neo4j.ExecuteQuery(ctx, driver,
    `MATCH (p:Person {name: $name})-[:KNOWS]->(friend)
     RETURN friend.name AS name`,
    map[string]any{"name": "Alice"},
    neo4j.EagerResultTransformer,
    neo4j.ExecuteQueryWithDatabase("neo4j"),       // always specify
    neo4j.ExecuteQueryWithReadersRouting(),         // for read queries
)
if err != nil {
    return fmt.Errorf("query people: %w", err)
}

for _, record := range result.Records {
    name, _ := record.Get("name")
    fmt.Println(name)
}
fmt.Println(result.Summary.Counters().NodesCreated())

Key options:

neo4j.ExecuteQueryWithDatabase("mydb")          // required for performance
neo4j.ExecuteQueryWithReadersRouting()           // route reads to replicas
neo4j.ExecuteQueryWithImpersonatedUser("jane")  // impersonate
neo4j.ExecuteQueryWithoutBookmarkManager()       // opt out of causal consistency

❌ Never concatenate user input into query strings. Always use map[string]any parameters.

Managed Transactions (Session-Based)

Use for lazy streaming (large result sets) or callback-level control.

session := driver.NewSession(ctx, neo4j.SessionConfig{
    DatabaseName: "neo4j", // always specify
    AccessMode:   neo4j.AccessModeRead,
})
defer session.Close(ctx)

result, err := session.ExecuteRead(ctx,
    func(tx neo4j.ManagedTransaction) (any, error) {
        res, err := tx.Run(ctx,
            `MATCH (p:Person) RETURN p.name AS name LIMIT $limit`,
            map[string]any{"limit": 100},
        )
        if err != nil {
            return nil, err
        }
        var names []string
        for res.Next(ctx) { // lazy — don't Collect() on large sets
            name, _ := res.Record().Get("name")
            names = append(names, name.(string))
        }
        return names, res.Err()
    },
)

❌ No side effects in callback — retried on transient failures.

ExecuteRead → replicas. ExecuteWrite → cluster leader.

Explicit Transactions

Use when transaction work spans multiple functions or requires external coordination.

session := driver.NewSession(ctx, neo4j.SessionConfig{DatabaseName: "neo4j"})
defer session.Close(ctx)

tx, err := session.BeginTransaction(ctx)
if err != nil {
    return err
}
if err := doPartA(ctx, tx); err != nil {
    tx.Rollback(ctx)
    return err
}
if err := doPartB(ctx, tx); err != nil {
    tx.Rollback(ctx)
    return err
}
return tx.Commit(ctx)

❌ Not auto-retried. Caller handles retry. Prefer managed transactions unless you need explicit control.

Error Handling

result, err := neo4j.ExecuteQuery(...)
if err != nil {
    var neo4jErr *neo4j.Neo4jError
    if errors.As(err, &neo4jErr) {
        slog.Error("database error", "code", neo4jErr.Code, "msg", neo4jErr.Msg)
    }
    var connErr *neo4j.ConnectivityError
    if errors.As(err, &connErr) {
        slog.Error("connectivity error", "err", connErr)
    }
    return fmt.Errorf("execute query: %w", err)
}

Helpers:

neo4j.IsNeo4jError(err)                // server-side Cypher/database error
neo4j.IsTransactionExecutionLimit(err) // managed tx retries exhausted

In managed tx callback: return error → driver retries if transient.

ConnectivityError at startup: check URI scheme, credentials, firewall.

Data Types

| Cypher | Go |

|--------|----|

| Integer | int64 |

| Float | float64 |

| String | string |

| Boolean | bool |

| List | []any |

| Map | map[string]any |

| Node | neo4j.Node |

| Relationship | neo4j.Relationship |

| Path | neo4j.Path |

| Date | neo4j.Date |

| DateTime | neo4j.Time |

| Duration | neo4j.Duration |

| null | nil |

// Typed extraction (v6+, preferred):
neo4j.GetRecordValue[string](record, "name")

// Manual extraction:
rawAge, ok := record.Get("age")
if !ok { return errors.New("missing 'age' field") }
age := rawAge.(int64) // Neo4j integers → int64

// Node access:
rawNode, _ := record.Get("p")
node := rawNode.(neo4j.Node)
name := node.Props["name"].(string)
labels := node.Labels // []string

❌ Always check ok from record.Get() before type-asserting — panics on missing key.

❌ After lazy for res.Next(ctx) loop, always check res.Err().

Key Patterns

Context — always propagate

ctx, cancel := context.WithTimeout(parentCtx, 5*time.Second)
defer cancel()
// pass ctx to all driver calls

context.Background() has no deadline — slow queries block indefinitely.

Batching Writes

// Bad: one transaction per record
for _, item := range items {
    neo4j.ExecuteQuery(ctx, driver, writeQuery, item, ...)
}

// Good: UNWIND batch in one transaction
neo4j.ExecuteQuery(ctx, driver,
    `UNWIND $items AS item
     MERGE (n:Node {id: item.id})
     SET n += item`,
    map[string]any{"items": items},
    neo4j.EagerResultTransformer,
    neo4j.ExecuteQueryWithDatabase("neo4j"),
)

Generic Helpers (v6+)

Prefer type-safe helpers over manual assertions:

// GetRecordValue[T] — extract + cast in one call
name, isNil, err := neo4j.GetRecordValue[string](record, "name")
// isNil=true when OPTIONAL MATCH returned null; err != nil when key absent or wrong type

// CollectTWithContext — map all records to a slice
people, err := neo4j.CollectTWithContext(ctx, result, func(record *neo4j.Record) (Person, error) {
    name, _, err := neo4j.GetRecordValue[string](record, "name")
    age, _, _   := neo4j.GetRecordValue[int64](record, "age")
    return Person{Name: name, Age: int(age)}, err
})

// SingleTWithContext — expect exactly one record (error if 0 or 2+)
person, err := neo4j.SingleTWithContext(ctx, result, func(record *neo4j.Record) (Person, error) {
    name, _, _ := neo4j.GetRecordValue[string](record, "name")
    return Person{Name: name}, nil
})

// GetProperty — typed property from Node or Relationship
node, _, _ := neo4j.GetRecordValue[neo4j.Node](record, "p")
nameVal, err := neo4j.GetProperty[string](node, "name")

Spatial Types

// 2D Cartesian (SRID 7203), 3D Cartesian (SRID 9157)
pt2d := neo4j.Point2D{X: 1.23, Y: 4.56, SpatialRefId: 7203}
pt3d := neo4j.Point3D{X: 1.23, Y: 4.56, Z: 7.89, SpatialRefId: 9157}

// 2D WGS-84 (SRID 4326), 3D WGS-84 (SRID 4979)
london := neo4j.Point2D{X: -0.118092, Y: 51.509865, SpatialRefId: 4326}
shard  := neo4j.Point3D{X: -0.0865, Y: 51.5045, Z: 310, SpatialRefId: 4979}

// Pass as parameter
result, err := neo4j.ExecuteQuery(ctx, driver,
    "CREATE (p:Place {location: $loc})",
    map[string]any{"loc": london},
    neo4j.EagerResultTransformer,
    neo4j.ExecuteQueryWithDatabase("neo4j"),
)

// Read from result — assert to Point2D or Point3D
raw, _ := record.Get("location")
if p2d, ok := raw.(neo4j.Point2D); ok {
    fmt.Printf("lon=%f lat=%f srid=%d\n", p2d.X, p2d.Y, p2d.SpatialRefId)
}

// Distance (same SRID only)
result, _ = neo4j.ExecuteQuery(ctx, driver,
    "RETURN point.distance($p1, $p2) AS distance",
    map[string]any{"p1": pt2d, "p2": neo4j.Point2D{X: 10, Y: 10, SpatialRefId: 7203}},
    neo4j.EagerResultTransformer, neo4j.ExecuteQueryWithDatabase("neo4j"),
)
dist, _ := result.Records[0].Get("distance")
fmt.Println(dist.(float64))

Always Specify Database

neo4j.ExecuteQueryWithDatabase("neo4j")    // in ExecuteQuery
neo4j.SessionConfig{DatabaseName: "neo4j"} // in sessions

Omitting costs a network round-trip per call to resolve home database.

Causal Consistency

ExecuteQuery manages bookmarks automatically — no action needed for sequential calls.

Cross-session (parallel workers): combine bookmarks explicitly — see references/repository-pattern.md.

Common Errors

| Error / Symptom | Cause | Fix |

|-----------------|-------|-----|

| ConnectivityError at startup | URI wrong / TLS mismatch / firewall | Check scheme (neo4j+s:// for Aura), credentials, port 7687 |

| ConnectivityError mid-run | Pool exhausted | Increase MaxConnectionPoolSize; check for leaked sessions |

| Panic on type assertion | record.Get() returned nil/wrong type | Use neo4j.GetRecordValueT or check ok first |

| res.Err() non-nil after loop | Network error mid-stream | Handle error; re-run transaction |

| Callback retried unexpectedly | Side effect inside managed tx | Move side effects outside callback |

| Context deadline exceeded | No timeout on context | Use context.WithTimeout |

| 0 results, query looks correct | Wrong DatabaseName | Always set DatabaseName in config |

| CALL IN TRANSACTIONS fails | Run inside managed tx | Use session.Run() (auto-commit) |

References

Load on demand:

references/advanced-config.md — connection pool tuning, custom address resolver, notification config, Bolt logging, auth options, URI scheme table
references/repository-pattern.md — repository wrapper pattern, cross-session causal consistency with bookmarks

WebFetch

| Need | URL |

|------|-----|

| Go driver manual | https://neo4j.com/docs/go-manual/current/ |

| API reference | https://pkg.go.dev/github.com/neo4j/neo4j-go-driver/v6/neo4j |

Checklist

[ ] One driver created at startup; shared across goroutines; defer driver.Close(ctx)
[ ] driver.VerifyConnectivity(ctx) called at startup
[ ] DatabaseName set in all SessionConfig / ExecuteQueryWithDatabase
[ ] context.WithTimeout used for production queries
[ ] map[string]any parameters used — no string interpolation
[ ] ExecuteQueryWithReadersRouting() on read-only ExecuteQuery calls
[ ] res.Err() checked after lazy for result.Next(ctx) loop
[ ] Type assertions guarded (use GetRecordValue[T] or check ok)
[ ] No side effects inside managed transaction callbacks
[ ] session.Run() used for CALL IN TRANSACTIONS / auto-commit queries
[ ] Sessions closed with defer session.Close(ctx)

How to use it

Copy the folder

Take neo4j-contrib/neo4j-driver-go-skill from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.