Cloudtrail Threat Detector
Analysis
This skill is a coherent, instruction-only AWS CloudTrail analysis guide that asks users to provide exported logs and explicitly avoids credentials or direct AWS access.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
"tools: claude, bash" and "This skill is instruction-only. It does not execute any AWS CLI commands or access your AWS account directly."
The frontmatter names bash, while the body frames AWS CLI commands as user-run export examples and says the skill should not access AWS directly. This is not suspicious, but users should treat the commands as manual setup steps.
Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.
"Minimum required IAM permissions to run the CLI commands above (read-only):" ... "Action": ["cloudtrail:LookupEvents", "cloudtrail:GetTrail", "logs:FilterLogEvents", "logs:GetLogEvents"], "Resource": "*"
The skill documents read-only AWS permissions a user may need to export CloudTrail and CloudWatch logs. This is purpose-aligned and disclosed, but users should notice that it involves AWS account visibility.
