gsd-orchestrator

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of the gsd-pi Node.js package to provide the gsd binary. This package is a resource provided by the skill's vendor, gsd-build.- [COMMAND_EXECUTION]: The skill orchestrates the gsd CLI tool via shell subprocess commands (e.g., gsd headless auto). This is the primary mechanism for its autonomous software development functionality.- [DYNAMIC_EXECUTION]: The skill generates project specifications (spec.md) and configuration files (answers.json) at runtime, which are then processed and executed by the gsd CLI tool.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided specifications to drive the autonomous development process.
  • Ingestion points: User-provided product specifications passed to gsd via the --context or --context-text flags in SKILL.md and workflows/build-from-spec.md.
  • Boundary markers: The CLI uses distinct flags (--context, --answers) to separate inputs, and the workflow suggests using structured markdown for specifications.
  • Capability inventory: The gsd CLI tool called by the skill has broad capabilities including file system access, network operations (via generated code), and git management.
  • Sanitization: The skill relies on the internal logic and guardrails of the gsd tool and the underlying LLM; it does not perform explicit sanitization of the input specification.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 04:16 AM