block-inventory

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • COMMAND_EXECUTION (MEDIUM): The skill executes shell commands (ls) and invokes a Node.js script located in a sibling skill directory (.claude/skills/block-collection-and-party/scripts/search-block-collection-github.js). This creates a dependency on the integrity of external scripts that are not contained within the skill itself.
  • PROMPT_INJECTION (MEDIUM): This skill is vulnerable to Indirect Prompt Injection (Category 8).
  • Ingestion points: Reads local directory names (blocks/*/) and potential documentation or comments within block code to determine purposes.
  • Boundary markers: Absent. Information is gathered and consolidated without clear delimiters or instructions for the agent to ignore embedded commands.
  • Capability inventory: Performs command execution (ls, node) and the results directly influence the 'content modeling' phase of the migration.
  • Sanitization: Absent. Data from the local filesystem is treated as trusted metadata, which could allow a malicious repository to influence agent behavior.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 01:36 PM