catalog-repair
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the local command
memory-hub catalog-repairvia the Bash tool to scan the system for inconsistencies. - [INDIRECT_PROMPT_INJECTION]: The skill processes the contents of
topics.mdand files within the.memory/directory, creating a surface where malicious content in those files could influence the agent's repair actions. - Ingestion points: Reads from
topics.mdand various files in the.memory/folder. - Boundary markers: There are no specific delimiters or instructions provided to the agent to ignore potential instructions embedded within the memory files.
- Capability inventory: The agent is authorized to execute
memory-hub(Bash) andmemory.write(file writing) to implement repairs. - Sanitization: The skill does not define any sanitization or validation logic for the data read from the memory files before the AI performs its 'self-healing' writes.
Audit Metadata