sdlc-orchestrator
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides templates for executing various developer CLI tools (Gemini, Copilot, Claude Code, Aider, Kimi, Cursor) via the Bash tool. These patterns involve interpolating user-controlled strings directly into command line arguments (e.g.,
gemini -p "PROMPT"). This relies on the agent platform's underlying shell escaping to prevent command injection. - [PROMPT_INJECTION]: The skill facilitates indirect prompt injection by design. It takes arbitrary user requests and forwards them as prompts to sub-agents. There are no boundary markers or instructions to sanitize or wrap this untrusted data to prevent a malicious user from hijacking the sub-agent's behavior.
- Ingestion points: User-provided task descriptions enter the workflow via the prompt substitution variables in SKILL.md.
- Boundary markers: None are present; user input is directly concatenated into the command strings.
- Capability inventory: The skill has the capability to execute arbitrary commands via the Bash tool to call external CLIs.
- Sanitization: No sanitization or input validation logic is described in the skill instructions.
Audit Metadata