projects-setup

Warn

Audited by Gen Agent Trust Hub on Jul 25, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute a shell script for project initialization. The path to this script is dynamically resolved at runtime through shell expansion and subshell execution.
  • Evidence: bash "${CLAUDE_PLUGIN_ROOT:-$(ls -td "$HOME"/.claude/plugins/cache/*/cogni-projects/*/ 2>/dev/null | head -1)}/scripts/portfolio-init.sh" "<portfolio-slug>" "<portfolio-name>" in SKILL.md.
  • [PROMPT_INJECTION]: The skill captures user-supplied values for portfolio identity and interpolates them directly into a shell command. This creates an indirect prompt injection surface where crafted user input could lead to unintended command execution.
  • Ingestion points: User inputs for 'portfolio name' and 'portfolio slug' defined in the workflow steps.
  • Boundary markers: None are specified to wrap or delimit the user input within the command.
  • Capability inventory: The skill utilizes the Bash tool to execute system scripts.
  • Sanitization: No logic is provided to sanitize or escape shell metacharacters in the user-provided strings.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 25, 2026, 01:51 PM
Security Audit — agent-trust-hub — projects-setup