business-e2e-context
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is configured to execute local commands including
bun run context:mapandgit logto access project journey data and check for file staleness. These actions are standard for developer context tools and are restricted to the local environment. - [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection due to its requirement to read and interpret project source code and HTML map files.
- Ingestion points: Data is ingested from the project repository and
references/business-e2e-map.html. - Boundary markers: The instructions do not prescribe specific boundary delimiters for the ingested content.
- Capability inventory: The skill can invoke shell commands and use the
diagramscapability for visualization. - Sanitization: No specific data sanitization or filtering procedures are mentioned in the skill instructions.
Audit Metadata