gsd-pipeline-orchestrator
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute system commands and the GSD command suite (gsd-*) to automate project phases, build processes, and containerized verification via Docker. These operations are consistent with the skill's documented purpose as an orchestrator for the GSD framework.\n- [DYNAMIC_EXECUTION]: The orchestrator dynamically selects and spawns specialized agents and other skills (e.g., appsec-security-orchestrator, codex@openai-codex) based on real-time classification of the project's technical requirements. This includes automated injection of platform-specific skills (e.g., Vercel, Supabase) into the planner agent's context to ensure architectural consistency.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes project-specific data to drive orchestration logic, creating a surface for potential indirect prompt injection.\n
- Ingestion points: Reads project metadata files (PROJECT.md, ROADMAP.md), configuration files (package.json), and repository structure to classify phases.\n
- Boundary markers: Employs XML-style tags within the skill instructions to delimit operational blocks.\n
- Capability inventory: Accesses powerful orchestration tools including Bash, Skill, and Agent to dispatch tasks.\n
- Sanitization: The orchestrator lacks explicit sanitization of ingested project content before using it to influence the composition of sub-agent calls and pipeline commands.
Audit Metadata