goal-prompt

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted conversation data and repository artifacts to generate instructions for a downstream agent. This creates a potential surface for indirect prompt injection if the source content contains malicious instructions intended to influence the generated Goal Prompt.
  • Ingestion points: The skill reads the current conversation and referenced repository artifacts as specified in the Workflow section (SKILL.md).
  • Boundary markers: The output template does not explicitly require the use of delimiters or boundary markers to wrap content ingested from the conversation.
  • Capability inventory: The skill is restricted to text generation and does not define internal capabilities for file writing, network access, or command execution that could be abused by an injection.
  • Sanitization: No explicit sanitization, filtering, or validation of the source conversation content is performed before it is compressed into the output prompt.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 12:37 AM
Security Audit — agent-trust-hub — goal-prompt