meta-knowledge-base-bootstrap

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface where external data can influence agent behavior.
  • Ingestion points: Untrusted data from web search results (outputs.ingest) is ingested into the agent context in the index step in SKILL.md.
  • Boundary markers: The index step lacks delimiters or explicit instructions to the agent to disregard embedded commands within the interpolated search results.
  • Capability inventory: The skill possesses capabilities including executing other skills (skill_exec), persisting data to the filesystem (memory_save), and generating spreadsheets (xlsx).
  • Sanitization: Although the skill employs xml_escape and truncate in the classify and memorize steps, search results are interpolated directly into the xlsx task description without sanitization.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 12:04 PM
Security Audit — agent-trust-hub — meta-knowledge-base-bootstrap