sb-flows
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill includes
extract-flows.sh, which executes a Python script via a heredoc to process and format route information. Additionally, theextract-app-graph.mjsNode.js template is used to perform deep-edge tracing and repo-local resolution of navigation idioms through static code analysis.\n- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves scanning untrusted project source code to generate navigation metadata, which creates a potential surface for indirect prompt injection.\n - Ingestion points: Project source directories (src/, app/, pages/) are scanned by extraction scripts.\n
- Boundary markers: Instructions in
SKILL.mdandflow-capture.mdexplicitly recommend manual validation of the generated graph data and reconciliation of extraction counts.\n - Capability inventory: The skill uses Bash, Python, and Node.js to read files, perform regular expression matching, and write JSON artifacts to the
.storybook/directory.\n - Sanitization: The scripts use regular expressions for data extraction and standard libraries for JSON serialization, minimizing the risk of executable content escaping into the output.
Audit Metadata