The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references standard development tools and libraries including xstate, @xstate/react, vitest, eslint-plugin-boundaries, and dependency-cruiser. These are well-known packages within the JavaScript ecosystem used for their intended purposes.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were detected. The skill provides architectural guidance and code templates but does not perform any dynamic execution of untrusted remote scripts.
[DATA_EXFILTRATION]: There are no network operations or file system accesses that would lead to data exfiltration. The skill focuses on client-side state management patterns.
[CREDENTIALS_UNSAFE]: No hardcoded credentials, API keys, or secrets were found. The documentation correctly identifies handles like 'credentials' as resources to be managed securely within encapsulated actors.
[PROMPT_INJECTION]: The instructions do not contain any patterns attempting to override agent behavior, bypass safety filters, or extract system prompts. The use of 'IMPORTANT' and 'CRITICAL' is strictly for technical emphasis.
[INDIRECT_PROMPT_INJECTION]: The skill provides robust patterns for handling external data, such as 'identity fencing' and 'authoritative convergence', which are security best practices for mitigating risks from untrusted external inputs.
[DYNAMIC_CONTEXT_INJECTION]: The skill does not use any pre-execution shell command syntax (!command).

xstate