post-optimizer

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for Indirect Prompt Injection (Category 8) because its primary workflow involves ingesting and acting upon untrusted data from the internet.
  • Ingestion points: In Step 1b ('Hotspot Scanning & Research'), the skill extracts keywords from user input and uses them to perform web searches. The resulting content from external websites is then processed to build 'research conclusions'.
  • Boundary markers: There are no boundary markers (like XML tags or delimiters) or specific instructions provided to the agent to ignore or isolate potential commands embedded within the retrieved search data.
  • Capability inventory: The skill utilizes a 'search tool' to fetch data. The generated output is then presented to the user, creating a chain where malicious instructions from a third-party website could influence the final generated content or the agent's behavior.
  • Sanitization: The instructions do not include any steps for sanitizing, validating, or filtering the content found during the research phase before it is interpolated into the agent's reasoning process.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 06:22 AM