team-lifecycle-v3
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a structured and secure orchestration framework for software development. It includes dedicated roles for security auditing (security-expert) and quality review (reviewer), which actively scan for vulnerabilities like SQL injection and hardcoded secrets in the generated output.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from user requests and external codebases. However, this risk is mitigated by the modular design and human-in-the-loop checkpoints.\n
- Ingestion points: roles/pipeline/analyst.md (topic content), roles/pipeline/tester.md (source code and tests).\n
- Boundary markers: Agent prompts utilize structured tags to separate instructions from data.\n
- Capability inventory: Includes file system access (Read, Write), command execution (Bash), and sub-agent management (Agent).\n
- Sanitization: The workflow incorporates a reviewer agent that evaluates outputs against security standards (e.g., OWASP Top 10).\n- [COMMAND_EXECUTION]: The skill executes local developer tools and test frameworks (e.g., ccw cli, npm, pytest) to perform its duties. These executions are confined to the local project environment and the vendor's provided CLI tools.
Audit Metadata