knowledge-harvester

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is a formatting tool with no malicious patterns, obfuscation, or unauthorized network activity detected. It requires explicit human approval for all file operations.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external research artifacts, creating a potential surface for indirect prompt injection.
  • Ingestion points: Research sources and synthesized connections cataloged during the workflow (SKILL.md).
  • Boundary markers: Mandatory human approval is required before any write operation to the vault (SKILL.md).
  • Capability inventory: The skill is limited to writing structured Markdown and YAML files to local storage (SKILL.md); it possesses no network, shell execution, or dynamic code loading capabilities.
  • Sanitization: All ingested content is placed within strict Markdown and YAML templates defined in the formatting guide (references/vault-format.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 12:42 AM
Security Audit — agent-trust-hub — knowledge-harvester