plantuml-skill

Pass

Audited by Gen Agent Trust Hub on May 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses curl to transmit diagram source code to https://kroki.io, a well-known public API used for rendering diagrams into various formats like PNG and SVG.\n- [COMMAND_EXECUTION]: Executes system commands via curl for network requests and git for repository management and update checks.\n- [REMOTE_CODE_EXECUTION]: Includes a self-update mechanism that uses git pull to fetch and apply updates from the vendor's official repository at https://github.com/Agents365-ai/plantuml-skill. The operation is designed to be user-controlled and requires explicit consent before execution.\n- [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it processes user-provided descriptions to generate executable diagram code which is then sent to an external service.\n
  • Ingestion points: User-provided diagram requests processed in SKILL.md.\n
  • Boundary markers: None present to differentiate user content from system instructions.\n
  • Capability inventory: Network access via curl, file system interaction, and shell command execution using git.\n
  • Sanitization: No sanitization or validation of the generated PlantUML source code is performed before transmission.
Audit Metadata
Risk Level
SAFE
Analyzed
May 13, 2026, 12:36 PM