Aws Service Chaos Research

AdvisoryAudited by Static analysis on May 10, 2026.

Overview

No suspicious patterns detected.

Findings (0)

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.

What this means

The skill may not work unless the AWS Knowledge MCP server and/or AWS CLI are already configured.

Why it was flagged

The skill relies on external MCP tools and optionally AWS CLI, while the registry metadata declares no required binaries, env vars, or credentials. This is disclosed in the docs and no automatic installer is present, so it is a compatibility/transparency note rather than a concern.

Skill content
Required tools (at least one of each group): ... `aws___read_documentation` ... AWS CLI `aws fis list-actions`
Recommendation

Use a trusted aws-knowledge-mcp-server setup and confirm which AWS CLI profile/region will be used before invoking the skill.

What this means

The agent may use local AWS CLI configuration and shell utilities to discover region-specific FIS actions.

Why it was flagged

The skill instructs the agent to run local shell/AWS CLI commands. These commands are read-only and directly support the stated purpose, but they still execute in the user's local environment.

Skill content
run `aws configure get region` ... `aws fis list-actions --region TARGET_REGION | jq '.actions[].id' | awk -F':' '{print $2}' | sort -u`
Recommendation

Confirm the intended AWS profile and region, and use least-privilege/read-only permissions for FIS action discovery.

What this means

If AWS CLI is configured, the skill may make AWS API calls under the user's account to list available FIS actions.

Why it was flagged

AWS CLI use normally relies on the user's configured AWS credentials/profile. The documented use is for FIS action discovery, not mutation, and there is no evidence of credential logging or unrelated account access.

Skill content
AWS CLI (`aws`) | FIS action discovery (preferred) | Falls back to documentation search if unavailable
Recommendation

Run with an AWS profile that has only the read permissions needed for discovery; do not grant experiment-creation or destructive permissions unless you separately intend to use them.

What this means

Documentation lookups may disclose the AWS service and region being researched to the configured MCP provider.

Why it was flagged

The skill sends documentation queries and page-read requests through AWS Knowledge MCP. This is disclosed and purpose-aligned, but service names, regions, and any user-provided context included in queries may leave the local agent boundary.

Skill content
All documentation research uses **only** the AWS Knowledge MCP tools above. Do NOT use SearXNG or other web search tools for documentation research.
Recommendation

Avoid putting secrets, account IDs, or sensitive resource identifiers into search phrases, and use a trusted MCP server.

What this means

Following the generated recommendations without safeguards could disrupt AWS workloads or dependent systems.

Why it was flagged

The report template includes potentially disruptive chaos-engineering actions. In context these are recommendations/hints for a chaos-testing report, not automatic execution instructions, but they could cause outages if a user later runs them in production.

Skill content
`aws:network:disrupt-connectivity` ... `aws:ec2:stop-instances` / `terminate-instances` ... `aws:ssm:send-command`
Recommendation

Treat the output as a planning document: require human approval, define stop conditions, limit blast radius, and test in controlled/non-production environments where possible.