skills/dcoon/agent-kit/wiki-schema/Gen Agent Trust Hub

wiki-schema

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill reads and processes user-provided notes from an Obsidian vault, which introduces a surface for indirect prompt injection if those notes contain malicious instructions.
  • Ingestion points: The skill utilizes note_read, periodic (to get daily notes), and multiple search operations (search_text, search_regex, search_metadata, search_semantic) to ingest data from the Obsidian vault.
  • Boundary markers: The skill contains several GUARD and HALT instructions that limit autonomous actions and protect human-authored content. It also defines a ## Provenance section for auditing changes. However, there are no specific markers or instructions to treat note content as untrusted data or to ignore embedded commands.
  • Capability inventory: The skill possesses extensive file manipulation capabilities within the vault, including note_create, note_write, note_patch, note_move, and note_delete operations.
  • Sanitization: There is no evidence of sanitization or validation of the note content before it is processed by the agent to create new entities or update existing records.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 12:27 AM