The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources that is stored in local files. The agent is instructed to use this data for generating business projections, which creates an attack surface where malicious content in the source data could influence agent behavior.
Ingestion points: The skill reads from references/economic-metrics.csv and references/economic-metrics-history.csv to identify trends and cite figures in its output.
Boundary markers: Absent. There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content within the CSV data files.
Capability inventory: The skill itself provides no tools, but the instructions assume a Python environment for the scrape_metrics.py tool. The execution environment (e.g., Claude Code) typically possesses file read/write and terminal execution capabilities.
Sanitization: Absent. The skill does not specify any validation, filtering, or sanitization steps for the data read from the CSV files.
[COMMAND_EXECUTION]: The skill documentation includes an explicit instruction to execute a local Python script to refresh reference data.
Evidence: The file SKILL.md contains the instruction: Run python tools/scrape_metrics.py kenya to refresh. This involves the execution of code shipped with the skill to perform network operations and file writes.

kenya