default-workflow

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and follows instructions from an external local file.
  • Ingestion points: The agent is instructed to read ~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md immediately upon activation in SKILL.md.
  • Boundary markers: No delimiters or 'ignore embedded instructions' warnings are present to isolate the workflow content.
  • Capability inventory: The skill allows the external workflow file to dictate the invocation of multiple specialized agents (e.g., architect, security, builder), granting it significant control over the agent's logical flow.
  • Sanitization: There is no evidence of sanitization or validation of the content read from the workflow file before it is used to direct agent behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 03:25 PM