cloud-audit-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| AWS_REGION | No | Region for EC2, RDS and CloudTrail calls in aws mode. | us-east-1 |
| AWS_PROFILE | No | AWS profile name, including SSO profiles, used by the standard AWS credential chain in aws mode. | |
| CLOUD_AUDIT_MODE | No | fixture serves JSON; aws queries the account. | fixture |
| AWS_ACCESS_KEY_ID | No | AWS access key ID used by the standard AWS credential chain in aws mode. | |
| CLOUD_AUDIT_FIXTURE | No | Path to an alternative estate JSON in fixture mode. | bundled estate |
| AWS_SECRET_ACCESS_KEY | No | AWS secret access key used by the standard AWS credential chain in aws mode. | |
| CLOUD_AUDIT_LOG_LEVEL | No | Log level: error, warn, info or debug. Logs go to stderr. | info |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| whoamiA | Caller identity, account id, region and whether the server is in fixture or aws mode. |
| list_checksB | The catalog of security-posture checks with id, title, severity, service and description. |
| run_checkA | Runs a single check by id and returns every finding it produced. |
| run_all_checksA | Runs the whole catalog, sorted worst-first, optionally filtered to a minimum severity. |
| describe_resourceA | Looks up one resource by ARN, id or name and returns its normalised shape plus the findings that reference it. |
| summarize_postureA | Counts findings by severity and service, lists the top five risks and scores the account 0–100. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| security_review | Guides the model through running every check and writing a prioritised remediation report. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| Check catalog | Every check with its metadata, as JSON. |
| Fixture estate | The sample estate used in fixture mode. |
TDQS
Scored across 6 tools
Each tool targets a distinct operation: identity (whoami), catalog (list_checks), single-check execution (run_check), full sweep (run_all_checks), resource lookup (describe_resource) and aggregate scoring (summarize_posture). The run_check vs run_all_checks pair is clearly separated by scope, and describe_resource is distinguished by operating on a resource rather than a check.
Five of six tools follow a clean snake_case verb_noun pattern (list_checks, run_check, run_all_checks, describe_resource, summarize_posture). The only deviation is whoami, which is a conventional single token and does not break readability.
Six tools is a tight, well-scoped surface for a security-posture audit server, with no redundant entries. Every tool maps to a distinct step in the audit workflow (identify, enumerate, run, inspect, summarize).
The core audit lifecycle is covered: discovery, execution, drill-down and aggregate reporting. Minor gaps remain — no way to list or search resources in bulk (only describe by ARN/id/name), no targeted finding query beyond check execution, and no export/report output.