The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze data from a 'Digital research engine' and external trade references, which constitutes an indirect prompt injection surface.
Ingestion points: Data ingestion occurs during the trade framework verification (Step 6) and the country evidence gathering described in the references.
Boundary markers: While the skill lacks technical delimiters (like XML tags) for external data, it explicitly instructs the agent to treat every country fact as a 'check' to be verified and recorded with a source and date.
Capability inventory: The skill's scope is restricted to analysis, drafting, and searching evidence; it does not authorize high-risk operations such as arbitrary command execution, file writing, or network exfiltration.
Sanitization: There are no automated sanitization mechanisms mentioned, with the skill relying instead on manual verification workflows and 'kill criteria' to handle unverified or failed country checks.
[NO_CODE]: The skill consists entirely of Markdown documentation and reference files, containing no scripts, executables, or code-based logic that could be exploited for direct execution attacks.

meta-international-market-entry