The Agent Skills Directory

[SAFE]: The skill is composed entirely of markdown instructions and templates for financial accounting. No malicious patterns, command injections, or exfiltration vectors were identified in the instructional content.
[NO_CODE]: The skill does not include any scripts, binaries, or configuration files that execute code. It relies solely on natural language processing within the agent's context.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external financial data such as sub-ledgers and bank statements, which constitutes an ingestion surface.
Ingestion points: Identified in the Inputs section of SKILL.md (e.g., trial balances, sub-ledger details, bank statements).
Boundary markers: Not explicitly defined in the output templates; however, the skill operates on structured financial data.
Capability inventory: No executable capabilities, shell tools, file-writing functions, or network access tools are included in the skill.
Sanitization: The skill includes a dedicated Input Sanitization section that mandates account validation and treats all external inputs as read-only artifacts, mitigating risks associated with data manipulation.

finance-reconciliation