The Agent Skills Directory

[SAFE]: The skill implements strong safety boundaries by including explicit legal disclaimers, referencing authoritative primary sources, and clearly defining the scope of its informational output to prevent the AI from attempting to prepare or sign official returns.
[EXTERNAL_DOWNLOADS]: The skill fetches data and provides links to official US government websites, including irs.gov, fincen.gov, and govinfo.gov, for the purpose of verifying tax thresholds and filing deadlines.
[COMMAND_EXECUTION]: Includes a bundled Python script, scripts/filing_map.py, designed to perform tax threshold calculations locally. The script takes user input as command-line arguments and does not engage in network activity or unauthorized file access.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user financial data to map filing obligations. While this represents a data ingestion surface, the risk is negligible as the skill lacks high-privilege capabilities such as data exfiltration or file writing.
Ingestion points: User-provided financial balances, filing status, and residency history collected in Stage 2 of the instructions.
Boundary markers: None explicitly defined in the prompt instructions.
Capability inventory: Limited to informational text generation and local math calculations via a bundled script.
Sanitization: No specific sanitization or filtering is performed on the ingested financial figures.

us-israel-dual-tax-navigator