The Agent Skills Directory

[PROMPT_INJECTION]: No attempts to bypass safety filters or override instructions were found. The workflow is strictly focused on business logic.
[DATA_EXFILTRATION]: The skill does not access sensitive credentials, configuration files, or local user data. It does not perform unauthorized network transmissions.
[OBFUSCATION]: Analysis of the skill's source files shows no use of Base64 encoding, zero-width characters, homoglyphs, or other techniques to hide content.
[REMOTE_CODE_EXECUTION]: No logic for downloading and executing external scripts or installing third-party packages was detected.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests external data for partner modeling, it provides strong instructions for verifying inputs and labeling unverified data, which reduces the surface for indirect injection attacks.
[DYNAMIC_CONTEXT_INJECTION]: The skill does not utilize shell-interpolated commands within its documentation or instructions.

partner-economics