resource-development

Pass

Audited by Gen Agent Trust Hub on May 10, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill uses a 'gangster' persona ('Underboss', 'The Don') to frame project management tasks. This is a stylistic role-play choice and does not attempt to override AI safety protocols or disregard instructions.
  • [PROMPT_INJECTION]: The workflow presents a surface for indirect prompt injection by processing a 'Contract' to generate verification commands.
  • Ingestion points: Decomposes an external 'Contract' to create task packages (SKILL.md).
  • Boundary markers: Absent; no delimiters or warnings to ignore instructions within the contract are specified.
  • Capability inventory: Includes file creation, git branch management, and shell command execution for task verification.
  • Sanitization: Absent; the skill does not explicitly validate or sanitize content from the 'Contract' before using it to generate executable commands.
  • [COMMAND_EXECUTION]: The skill instructs the agent to create and run shell commands for verifying work packages. This is standard functionality for a development automation tool and does not include hardcoded malicious commands or privilege escalation attempts.
  • [DATA_EXFILTRATION]: No network operations, exfiltration patterns, or attempts to access sensitive credentials (e.g., SSH keys, environment variables) were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
May 10, 2026, 02:14 PM
Security Audit — agent-trust-hub — resource-development