soia-pkm-distill

Pass

Audited by Gen Agent Trust Hub on Jul 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions specify an installation command using npx to fetch and execute code from the author's repository (soia-team/soia-open-skills). Additionally, it executes file system operations to read, edit, and create notes within the Obsidian vault.\n- [EXTERNAL_DOWNLOADS]: During setup, the skill downloads components from the soia-team/soia-open-skills repository, which is a resource provided by the skill author.\n- [REMOTE_CODE_EXECUTION]: The skill utilizes npx to add its functionality from a remote repository, a pattern that involves remote code acquisition and execution at the time of installation.\n- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting and processing untrusted content from user notes in the Obsidian vault.\n
  • Ingestion points: Reads article content from the user's <vault-articles-dir>/ as defined in SKILL.md.\n
  • Boundary markers: The instructions lack specific delimiters or warnings to prevent the agent from following potential commands embedded within the notes being processed.\n
  • Capability inventory: The skill is authorized to read, write, and create files on the local file system to manage the Obsidian vault content.\n
  • Sanitization: There is no evidence of sanitization, filtering, or validation of the input text from the vault articles before it is included in the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 11, 2026, 03:07 AM
Security Audit — agent-trust-hub — soia-pkm-distill