dream-memory
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data, specifically session transcripts and system logs, to update the agent's memory files. This creates a risk where malicious instructions embedded in those logs could influence the consolidation process.
- Ingestion points: The prompt template in
references/prompt-template.mdand the workflow inSKILL.mddirect the agent to review external logs and transcripts for durable facts. - Boundary markers: There are no explicit delimiters or instructions to treat the ingested log content as untrusted data.
- Capability inventory: The skill grants the agent the ability to read existing memory files and write updated topic files to the file system via the workflow and the helper script
scripts/dream_memory.py. - Sanitization: No sanitization or validation of the input content is performed before it is processed by the agent.
Audit Metadata