The Agent Skills Directory

[DATA_EXFILTRATION]: The skill performs broad discovery by reading the contents of sensitive configuration and environment files.
Evidence: Grep commands target .env, .env.local, .env.production, and .env.staging files to identify ports and URLs.
Evidence: cat commands are used to read .mcp.json and .claude/mcp.json files.
Security Concern: Environment files and MCP configurations are standard locations for storing authentication tokens, database credentials, and secret keys. Reading these files exposes sensitive information to the agent's context.
[COMMAND_EXECUTION]: The skill utilizes shell commands for project discovery and configuration.
Evidence: It uses find, grep, cat, and ls to search for files and read content to understand the project structure.
Evidence: It uses npx as a tool for setting up analytics infrastructure.
[EXTERNAL_DOWNLOADS]: The skill integrates external functionality by setting up MCP servers from the npm registry.
Specifically, it includes steps to fetch the @mixpanel/mcp-server and @posthog/mcp-server packages using npx to initialize analytics integration.
These packages provide integration with well-known analytics services.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the analysis of project data.
Ingestion points: Files including test specifications (.spec.ts, .feature), source code (.js, .rb, .py), and route definitions are read to extract flow logic.
Boundary markers: Absent. The skill does not define delimiters or provide instructions to ignore malicious content when parsing files.
Capability inventory: File system access, shell command execution, and configuration of network-enabled tools.
Sanitization: Absent. Content parsed from the project files is directly used to generate descriptive fields in YAML flow definitions.

flowchad-setup