anti-amnesia
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to process and report on previous tool results and conversation history. This establishes a processing surface for untrusted data that may have entered the context in prior turns.
- Ingestion points: SKILL.md (Instructions to use tool results, assistant responses, and receipts as the primary record).
- Boundary markers: Absent; there are no specific markers or instructions to ignore embedded commands within the historical record.
- Capability inventory: None; the skill consists only of markdown and YAML configuration with no associated scripts or tool definitions.
- Sanitization: Absent; the skill does not specify any sanitization or validation of the recorded data before it is summarized for the user.
- [SAFE]: The skill is entirely instructional and promotes transparency through the use of 'handoff receipts'. It does not perform network operations, access sensitive files, or execute commands.
Audit Metadata