The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch live documentation and search for integration patterns from external sources such as NPM, GitHub, and Stripe's documentation. While these are reputable sources, data fetched via WebFetch and WebSearch is technically untrusted and could be manipulated to include hidden instructions that influence the agent's behavior during code generation.
Ingestion points: External fetches to npmjs.com, docs.stripe.com, and results from WebSearch are directed in the 'Before writing code' section of SKILL.md.
Boundary markers: The instructions lack delimiters or specific warnings to the agent to disregard instructions potentially embedded within the fetched external data.
Capability inventory: The skill has access to powerful tools including Write, Edit, and Bash, which are used to implement server-side middleware and configuration files.
Sanitization: There are no instructions provided to sanitize, filter, or validate the content retrieved from the web before the agent incorporates it into its code-writing tasks.

mpp-server-middleware