How to Configure Goose DynamoDB Audit Log
To set up a goose dynamodb audit log, turn on cloud trail data events for your tables. Then bind your AI agent to a dedicated cloud role so every table call carries a unique tag. Engineering teams often run Goose with local cloud profiles to inspect items and fix bugs. When those local profiles hold broad permissions, the agent can read or change any item in your tables. How to set up a goose dynamodb audit log A native audit setup relies on your cloud provider logging service and identity ru
Your coding agents run with your access.
Claude Code reaches prod with a developer's credentials, and approval prompts get clicked through. hoop.dev checks every query, masks sensitive data, and holds risky changes for a human.
To set up a goose dynamodb audit log, turn on cloud trail data events for your tables. Then bind your AI agent to a dedicated cloud role so every table call carries a unique tag. Engineering teams often run Goose with local cloud profiles to inspect items and fix bugs. When those local profiles hold broad permissions, the agent can read or change any item in your tables.
How to set up a goose dynamodb audit log
A native audit setup relies on your cloud provider logging service and identity rules. First, you create a dedicated cloud role just for the Goose agent. You do not let the agent share a personal developer profile on a laptop. Instead, the developer assumes this special agent role before starting a work session.
Next, you attach a permission policy to that role for your specific DynamoDB tables. This policy lists the exact read or write actions the agent can run. You then set up your cloud logging trail to capture data events on those tables. Management events only show table creation or deletion, so data events are required here.
When Goose runs a command or script, the cloud API checks the active role. The logging service records the role name, the timestamp, the source IP, and the table name. Your security team then sends these log files to a central storage bucket for review.
What native table logging solves and what it misses
Native cloud logs prove which role called a table and when the call happened. They help your team spot unauthorized role usage or weird spikes in table reads. If an agent tries to call a blocked table, the log records a denied request.
Still, native cloud logs only see the outer API request at the cloud boundary. They decide who can log in and which broad API actions a role can run. They do not inspect the actual item values returned inside the payload in flight. If a table holds emails, phone numbers, or card numbers, the agent receives the raw values. Data masking rewrites matched fields in results so raw sensitive strings do not reach the prompt.
Audit Log Integrity + DynamoDB Fine-Grained Access: architecture patterns and best practices
Continue reading? Get the full guide.
Free. No spam. Unsubscribe anytime.
Native logs also struggle to tie a specific human approval to a single destructive write. Once a session has write permissions, the agent can update or delete thousands of items fast. Cloud trails also deliver log files in batches with a delay of several minutes. By the time a security alert fires, the bad table write is already done.
Managing agent sessions and data privacy in flight
To close these gaps, teams need controls that work on each individual request in real time. Rather than giving an agent static keys on a laptop, teams isolate credentials from the client. Each table query or update should link directly to an individual human user identity.
When an agent attempts a risky write action, the workflow should pause for human review. A teammate can check the exact command and approve it from their chat app. Also, filtering data in flight rewrites emails and card numbers before the agent processes the response. This keeps sensitive customer fields out of external model prompts and local session files.
Check database and API queries on the wire
Identity tools, roles, and network rules decide who opens a session. Setup alone cannot check what each query does or mask sensitive values in a response once a session starts. When workflows touch Postgres, MySQL, SQL Server, MongoDB, ClickHouse, Kubernetes APIs, SSH, or internal HTTP APIs, hoop.dev runs beside those services as an open-source proxy, a sidecar for your Postgres or API. It inspects requests coming in and results going back on the wire.
Your login details and client tools stay as they are while hoop.dev guardrails refuse bad deletes or drops before they run. On responses, hoop.dev data masking rewrites matched sensitive values in memory. Furthermore, hoop.dev holds SOC 2 Type II, is MIT licensed, and writes every query and decision to an audit trail that generates evidence for HIPAA, GDPR, and PCI DSS. See the source on GitHub.
Frequently asked questions
Does DynamoDB record item reads and writes by default?
No, default cloud trails only record table management actions like creating or deleting a table. You must turn on data events in your trail settings to log item reads and writes.
Do cloud audit logs show the actual data values returned to Goose?
No, cloud data event logs record the API action, role, and item keys. They do not record the full item payload or mask sensitive fields like emails or card numbers.
Why should Goose use a separate role instead of a developer profile?
Sharing a developer profile mixes human actions and agent actions in your audit logs. A separate role lets you track agent activity and restrict the agent to specific tables.
Related guides
In this post