gds-brainstorm-game

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Python script using 'uv run' to resolve workflow customization settings and process finalization tasks. These executions are triggered during activation in SKILL.md and upon completion in step-04-complete.md.
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates a brainstorming session that ingests external data from the user and from configuration files, which presents a surface for indirect prompt injection.
  • Ingestion points: User-provided brainstorming ideas, concepts, and themes are collected in step-03-ideation.md.
  • Boundary markers: The workflow does not explicitly define delimiters or instructions to ignore potential commands embedded in the user's ideation input.
  • Capability inventory: The skill has the capability to execute local scripts via 'uv run' (SKILL.md, step-04-complete.md) and perform file write operations to the session output file (step-01-init.md, step-02-context.md, step-03-ideation.md, step-04-complete.md).
  • Sanitization: No explicit sanitization or validation logic is applied to the user's input before it is written to the output document or used to influence the workflow state.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 09:39 PM
Security Audit — agent-trust-hub — gds-brainstorm-game