confluence-distill

Pass

Audited by Gen Agent Trust Hub on Apr 9, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [DATA_EXFILTRATION]: The skill makes network requests to well-known services including Atlassian (atlassian.net) and Membrane (skills.sh) to retrieve page data. These operations are essential for the distillation workflow and use legitimate endpoints.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) due to its core function of ingesting untrusted content from external Confluence pages to generate new instructions. \n
  • Ingestion points: Page bodies (body.storage/body.view) are retrieved from Confluence and processed in Step 3 of SKILL.md.\n
  • Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore' instructions to isolate the fetched external data from the agent's internal logic.\n
  • Capability inventory: The skill possesses the ability to create new directories and files on the host system and perform network requests to read content.\n
  • Sanitization: No sanitization or verification of the fetched content is required before it is used to write the root SKILL.md and reference files for the new package.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 9, 2026, 07:20 AM