The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill configures a local Express server that ingests JSON payloads and URL parameters. If an agent interacts with this server to manage diagram data, it could be exposed to malicious instructions embedded within the mock or proxied data.
Ingestion points: The express.json() middleware in src/dev/server.ts and route parameters (:id, :pageId) in src/dev/mocks.ts within SKILL.md accept external data.
Boundary markers: Absent; the instructions do not provide delimiters or specific warnings to the agent regarding the untrusted nature of the data returned by the dev server.
Capability inventory: The skill allows the use of Read, Write, and Edit tools as defined in the YAML frontmatter of SKILL.md.
Sanitization: The provided code snippets lack input validation or sanitization for the data processed by the mock routes or proxy middleware.
[SAFE]: The skill installs well-known and widely used Node.js packages (express, axios, vitest, etc.) from the standard registry for their intended purposes in a development environment.
[SAFE]: Environment variable management follows standard practices by using .env.example and .env files for local credential configuration.

lucidchart-local-dev-loop