The Agent Skills Directory

[SAFE]: The skill uses deterministic Python scripts that rely exclusively on the standard library. No external dependencies, network operations, or sensitive file system accesses were found across scripts/partner_tier_classifier.py, scripts/joint_gtm_planner.py, or scripts/revshare_modeler.py.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where it ingests data from a user-populated JSON file (generated from assets/partnership_intake_template.md). However, this is documented behavior, and the processing scripts perform validation and deterministic scoring without executing input as code or passing it to sensitive system calls. The risk is minimal as the scripts do not rely on LLM interpolation for their primary logic.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were identified. All scripts are local to the skill package and do not fetch external content.

partnerships-architect