default-workflow
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill implements a 'thin wrapper' architectural pattern where core execution instructions are dynamically loaded from a local file at
~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md. This constitutes an indirect prompt injection surface as the agent's behavior is governed by data ingested from the local file system. - Ingestion points:
~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md(read via theReadtool). - Boundary markers: Absent; the skill explicitly directs the agent to 'Follow all steps exactly' as specified in the external file.
- Capability inventory: The workflow involves file system modifications, GitHub PR and issue management, versioning, and code implementation.
- Sanitization: Absent. However, as this file is part of the vendor-owned application configuration directory, the risk is consistent with standard application functionality.
- [SAFE]: No evidence of hardcoded credentials, data exfiltration, or unauthorized network operations was detected. The skill's file access is limited to its own configuration directory.
- [NO_CODE]: This skill consists entirely of instructional markdown and does not contain any executable scripts or binary files.
Audit Metadata