The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the ecc-agentshield package via npm and references a third-party GitHub repository (affaan-m/agentshield). These resources are not from a well-known or verified vendor, presenting a potential supply chain risk.
[REMOTE_CODE_EXECUTION]: The skill frequently uses npx to execute code directly from the npm registry (npx ecc-agentshield scan). Running unverified code from a remote registry is a significant security risk as the content of the package can change at any time.
[COMMAND_EXECUTION]: The skill performs several shell-based operations, including scanning the file system and initializing configurations. Notably, the --fix flag allows the external tool to automatically modify local configuration files, which could lead to unauthorized changes if the tool is compromised.
[PROMPT_INJECTION]: The skill processes untrusted project data that may contain malicious instructions designed to influence the scanner or the agent's behavior.
Ingestion points: Reads sensitive project files including .claude/settings.json, CLAUDE.md, mcp.json, and various files within the hooks/ and agents/ directories.
Boundary markers: The instructions do not define delimiters or specific warnings to ignore embedded instructions within the processed data.
Capability inventory: The skill possesses the capability to execute shell commands and modify the local file system.
Sanitization: There is no mention of sanitization, validation, or filtering of the content being scanned.

security-scan