grill-with-docs
Pass
Audited by Gen Agent Trust Hub on Jul 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to ingest external data from a 'knowledge vault' (referenced in
docs/agents/knowledge-vault.md) using theresearch-from-vaultcapability. This creates a surface for indirect prompt injection where untrusted content in the vault could influence the agent's actions. - Ingestion points: Data is pulled from vault sources via
research-from-vaultor directly from the vault configuration. - Boundary markers: The skill does not specify any delimiters or instructions to ignore embedded commands within the vault data.
- Capability inventory: The skill has the capability to generate and write files, specifically Architecture Decision Records (ADRs) and glossaries.
- Sanitization: No sanitization or validation of the ingested vault content is described.
Audit Metadata