The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill recommends executing code from remote sources using 'npx'. This includes the 'cc-company-dashboard' package (a vendor-provided resource) and third-party packages such as '@cocal/google-calendar-mcp'. Executing unpinned packages from public registries poses a risk of supply chain attacks.
[COMMAND_EXECUTION]: The skill provides complex shell commands for the configuration of Model Context Protocol (MCP) servers. These commands include environment variables that point to sensitive local files, such as Google OAuth credentials ('GOOGLE_OAUTH_CREDENTIALS=/path/to/credentials.json').
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8). It collects raw user responses regarding business goals and activities and interpolates them directly into the '.company/CLAUDE.md' file, which is later used to define the agent's operational rules.
Ingestion points: User input collected during onboarding questions Q1 and Q2 in 'SKILL.md'.
Boundary markers: Absent; the template in 'references/claude-md-template.md' does not use delimiters or warnings to isolate user-provided content.
Capability inventory: The skill possesses file system read/write capabilities and generates shell commands for tool configuration.
Sanitization: The skill does not sanitize or escape user input before including it in the generated instruction files.
[EXTERNAL_DOWNLOADS]: The 'npx' execution path for the dashboard and MCP servers involves downloading external code from the npm registry at runtime.

company