AWS Bedrock Claude excessive use of tokens

name: AWS Bedrock Claude excessive use of tokens
id: a839a6f7-aaac-438b-9d99-be0b49481e17
version: 1
creation_date: '2026-07-06'
modification_date: '2026-07-06'
author: Rod Soto
status: production
type: Anomaly
description: Detects identities generating anomalously large model responses relative to their own historical baseline. For each identity, computes the average, maximum, and standard deviation of output token counts across all invocations, then flags any identity whose single largest response exceeds two standard deviations above their own mean. A statistically significant output spike from a single identity may indicate bulk data extraction, successful prompt injection producing verbose output, or a runaway agentic loop hitting context limits.
data_source:
    - AWS Bedrock Claude
search: |-
    `aws_bedrock_claude`
    | spath output="out_tokens" path="output.outputBodyJson.usage.output_tokens"
    | eval user = replace('identity.arn', ".*/", "")
    | stats count           AS invocations,
            avg(out_tokens) AS avg_out,
            max(out_tokens) AS max_out,
            stdev(out_tokens) AS stdev_out
      BY user, identity.arn
    | eval stdev_out  = coalesce(stdev_out, 0)
    | eval threshold  = avg_out + (2 * stdev_out)
    | where max_out > threshold
    | table
        user,
        identity.arn,
        invocations,
        avg_out,
        max_out,
        stdev_out,
        threshold
    | sort -max_out
    | `aws_bedrock_claude_excessive_use_of_tokens_filter`
how_to_implement: You must install and configure the Splunk Add-on for AWS (https://splunkbase.splunk.com/app/1876). Enable Amazon Bedrock model invocation logging in AWS so that Claude request/response payloads are delivered to S3 and/or CloudWatch Logs (see https://docs.aws.amazon.com/bedrock/latest/userguide/model-invocation-logging.html for setup steps), then ingest those logs into Splunk via the AWS TA. Configure the `aws_bedrock_claude` macro to point to the index and sourcetype (`json_no_timestamp`) where these logs land.
known_false_positives: This detection may produce false positives for identities with low invocation history, legitimate large document summarization tasks, or automated pipeline sessions with naturally variable output token counts.
references:
    - https://aws.amazon.com/blogs/apn/unlocking-the-power-of-splunk-with-amazon-bedrock-an-agentic-ai-approach-to-build-customized-splunk-assistants-using-bedrock-agents/
    - https://help.splunk.com/en/splunk-observability-cloud/observability-for-ai/splunk-ai-infrastructure-monitoring/set-up-ai-infrastructure-monitoring/amazon-bedrock
    - https://research.splunk.com/stories/aws_bedrock_security/
    - https://docs.aws.amazon.com/bedrock/latest/userguide/model-invocation-logging.html
drilldown_searches:
    - name: View the detection results for - "$user$"
      search: '%original_detection_search% | search user="$user$"'
      earliest_offset: $info_min_time$
      latest_offset: $info_max_time$
    - name: View risk events for the last 7 days for - "$user$"
      search: '| from datamodel Risk.All_Risk | search normalized_risk_object IN ("$user$") | stats count min(_time) as firstTime max(_time) as lastTime values(search_name) as "Search Name" values(risk_message) as "Risk Message" values(analyticstories) as "Analytic Stories" values(annotations._all) as "Annotations" values(annotations.mitre_attack.mitre_tactic) as "ATT&CK Tactics" by normalized_risk_object | `security_content_ctime(firstTime)` | `security_content_ctime(lastTime)`'
      earliest_offset: 7d
      latest_offset: "0"
intermediate_findings:
    entities:
        - field: user
          type: user
          score: 20
          message: Identity $user$ generated anomalously high output tokens in AWS Bedrock Claude with max output $max_out$ above threshold $threshold$.
analytic_story:
    - Suspicious AWS Bedrock Claude Activities
asset_type: Web Application
mitre_attack_id:
    - T1055
product:
    - Splunk Enterprise
    - Splunk Enterprise Security
    - Splunk Cloud
category: application
security_domain: endpoint
tests:
    - name: True Positive Test
      attack_data:
        - data: https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/aws_bedrock_claude/aws_bedrock_claude_excessive_use_of_tokens.ndjson
          sourcetype: json_no_timestamp
          source: http:bulkawsbedrock
      test_type: unit