scaffold-tracker-workflow

Pass

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The read-config.sh helper executes a hardcoded Python script for strict YAML parsing via python -c. This logic is internal to the platform resolution process and does not process untrusted code.
  • [COMMAND_EXECUTION]: Several components use execFileSync and spawnSync to interact with gh, git, and bash. These calls are restricted to well-known binaries and use controlled arguments to perform project probes and platform resolution.
  • [INDIRECT_PROMPT_INJECTION]: Untrusted data from external sources (Jira/GitHub column and status names) is ingested and written to the tracker-workflow.yaml file. Although this presents a potential injection surface, it is mitigated by sanitization logic.
  • Ingestion points: probeJira and probeGithub functions in scripts/scaffold-tracker-workflow.js which fetch live board data.
  • Boundary markers: Values are wrapped in single quotes in the emitted YAML output.
  • Capability inventory: File system write access (fs.writeFileSync) is used to save the configuration.
  • Sanitization: Single quotes within column names are escaped to double single-quotes ('') using replace(/'/g, "''") within the yamlScalar function in scripts/scaffold-tracker-workflow.js.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 10, 2026, 03:29 PM