The Agent Skills Directory

[SAFE]: No malicious commands, obfuscation, or unauthorized file system operations were detected.
[SAFE]: The skill uses designated tools (transaction_search, categorize) to perform financial data processing within its intended scope.
[PROMPT_INJECTION]: The skill's primary function involves processing external transaction descriptions, which represents a surface for indirect prompt injection (Category 8).
Ingestion points: Transaction descriptions are ingested via the transaction_search tool (SKILL.md).
Boundary markers: No specific delimiters or instructions to ignore embedded commands are used for transaction strings.
Capability inventory: The skill uses tools to read and modify transaction records but lacks higher-privilege capabilities like shell access or network requests.
Sanitization: Input is matched against patterns for categorization purposes; no specific sanitization is mentioned.
[SAFE]: Because the processing of descriptions is essential to the skill's intended accounting purpose, this surface is considered low risk and does not escalate the security verdict.

smart-categorize