workflow-from-chats
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process historical chat transcripts, which are considered untrusted data sources.
- Ingestion points: The skill explicitly instructs the agent to read both parent and subagent transcripts from the last 7 days (
SKILL.md). - Boundary markers: The instructions include guidance to cite parent conversations and filter anecdotes, but lack explicit delimiters or instructions to ignore potential adversarial commands embedded in the processed transcripts (Absent).
- Capability inventory: The skill is capable of generating persistent artifacts including new skills, rules, and workflow documents, which can influence or define future agent behavior (
SKILL.md). - Sanitization: There are no requirements for sanitizing or escaping the content extracted from transcripts before it is interpolated into new rules or skills (Absent).
Audit Metadata