investigation-workflow

Pass

Audited by Gen Agent Trust Hub on Nov 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the 'amplihack' CLI tool to run local workflow recipes, specifically 'amplifier-bundle/recipes/investigation-workflow.yaml'. This is a standard and expected functionality for managing complex workflows within the vendor's ecosystem.
  • [INDIRECT_PROMPT_INJECTION]: As the skill is designed to analyze and summarize external codebases and system architectures, it possesses an inherent risk of indirect prompt injection if the analyzed data contains malicious instructions. This risk is acknowledged as part of the tool's primary function.
  • [INDIRECT_PROMPT_INJECTION]: Mandatory Evidence Chain:
  • Ingestion points: Analyzes and explores external codebases, documentation, and system architectures (SKILL.md).
  • Boundary markers: Includes 'confirmation_required: true' in the metadata to ensure human oversight before executing key steps (SKILL.md).
  • Capability inventory: Includes the ability to execute shell commands via 'amplihack' and update local files like '.claude/context/PATTERNS.md' (SKILL.md).
  • Sanitization: No specific input sanitization for analyzed content is described in the instruction set, relying on the user confirmation gate.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 2, 2026, 04:52 PM