plugin-workflow
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted or project-specific data from multiple files and passes it to subagents, creating a surface for indirect prompt injection.
- Ingestion points: The skill reads
architecture.md,plan.md,creative-brief.md, andparameter-spec.mdas documented inreferences/stage-2-dsp.mdandreferences/stage-3-gui.md. - Boundary markers: Absent. The content of these files is interpolated directly into the subagent prompts (e.g., in the
constructPhasePromptlogic) without delimiters or instructions to ignore embedded directives. - Capability inventory: The orchestrator can invoke other agents via the
Tasktool and execute shell commands via theBashtool, providing a path for injected instructions to reach high-privilege operations. - Sanitization: Absent. There is no evidence of filtering or escaping the contents of the ingested files before they are sent to the subagents.
- [COMMAND_EXECUTION]: The skill performs significant local command execution to manage the build system, state, and file synchronization.
- Evidence: The skill executes several local scripts including
.claude/utils/sync-brief-from-mockup.sh(inreferences/creative-brief-sync.md),./scripts/build-and-install.sh(referenced inreferences/stage-1-foundation-shell.md), and numerousgit,sed, andshasumcommands for state persistence and auditing inreferences/state-management.md.
Audit Metadata