confluence

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and displays content from external Confluence pages, creating a surface for indirect prompt injection where malicious instructions embedded in documentation could influence the agent's actions.\n
  • Ingestion points: Page content is retrieved in scripts/confluence.py via the get_page and search_content functions.\n
  • Boundary markers: The output provided to the agent does not include explicit delimiters or specific warnings to ignore instructions contained within the fetched documentation.\n
  • Capability inventory: The skill has the capability to create and update pages or spaces and is permitted to run the Python integration script via the Bash tool.\n
  • Sanitization: While the skill performs Markdown and XHTML conversion, it does not implement specific sanitization to filter out adversarial prompt instructions from the page body.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 11:07 AM
Security Audit — agent-trust-hub — confluence