The Agent Skills Directory

[SAFE]: The skill consists entirely of text-based instructions for generating outreach sequences and reply handling. It does not contain executable code, shell commands, or external dependencies.
[REMOTE_CODE_EXECUTION]: No remote scripts, package installations, or external code sources are referenced in the skill files.
[DATA_EXFILTRATION]: There are no patterns indicating network operations or access to sensitive system files. The instructions explicitly warn against sending or uploading data without approval.
[PROMPT_INJECTION]: The instructions focus on establishing selection signals and maintaining compliance; there are no attempts to override agent behavior or bypass safety filters.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided signals for personalization, it does not involve tool outputs or automated workflows that would create a significant attack surface for indirect injection.

write-outreach