renewal-save-plan
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from various sources.
- Ingestion points: Defined in references/input-rules.md, the skill accepts conversational dumps, messy notes, Slack pastes, CRM tables, and PDF documents.
- Boundary markers: Section 0 of SKILL.md includes a mandatory instruction block that directs the agent to treat fetched or uploaded content as data, never as instructions.
- Capability inventory: The skill utilizes platform-native tools like web_search and web_fetch for factual verification but does not specify file-write or arbitrary subprocess execution within its logic.
- Sanitization: The skill relies on instructional boundaries and the agent's inherent safety filters to ignore embedded commands in the processed data.
Audit Metadata