peer-feedback
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data retrieved from communication tools, which presents a surface for indirect prompt injection attacks.
- Ingestion points: The skill uses
references/workiq-query-playbook.mdto fetch direct messages (Q1), shared group chats (Q2), mentions (Q3), and email threads (Q5). - Boundary markers: There are no explicit delimiters or instructions to the LLM to ignore or sanitize embedded instructions within the ingested chat or email data.
- Capability inventory: The skill uses the ingested data to generate a structured markdown feedback draft. No high-privilege capabilities such as arbitrary file writing, network exfiltration, or command execution were identified.
- Sanitization: The instructions do not define any sanitization, filtering, or escaping logic for the content retrieved from the external communication sources before it is interpolated into the generation prompt.
Audit Metadata