scenario
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources such as project files and user-supplied arguments to define its mission and scope.
- Ingestion points: The skill reads
$ARGUMENTSin the Mission section and project files includingdocs/product-roadmap.mdanddocs/project-config.jsonin Phase 0. - Boundary markers: No delimiters or explicit instructions are provided to the agent to separate untrusted data from the skill's own instructions or to ignore embedded commands.
- Capability inventory: The skill is designed to read from and write to the project's directory and is authorized to use
spawn_agentsub-agents for executing its protocol. - Sanitization: The ingested content from project artifacts and user inputs is processed without prior validation or sanitization.
Audit Metadata