Pharma GEO Playbook

What It Solves

Pharma GEO workflows often depend on protected platform capabilities, but local Python setup, authorization state, and cloud-side policy can still allow an Agent to continue generating results after a failed check. This skill puts Python environment detection, authorization gating, and protected-instruction validation before any business execution, reducing the risk of silent fallback when AUTH_REQUIRED, AUTH_INVALID, or remote authorization checks fail.

How It Works

• Detect Python first: before running scripts, it silently checks for a usable interpreter and records $PYTHON_CMD. If detection fails, it reports the issue instead of continuing. On Windows, python may open Microsoft Store; use py -3 or disable app execution aliases.
• Complete authorization through the script: scripts/auth.py handles local credential cache, WorkBuddy user binding, platform validation, and error-signal output. It avoids manual JWT parsing or direct handling of AppKey / AppSecret.
• Fetch protected instructions: after AUTH_OK, it calls scripts/protected.py to retrieve cloud-side protected instructions. Business logic may continue only after those instructions are returned successfully.
• Report failures explicitly: when signals such as CREDENTIALS_MISSING, CREDENTIALS_INVALID, HTTP 401, or network error appear, it stops business inference and tells the user what was detected and what to do next.

Notes

This skill is best treated as a protected-capability execution template: authorization and business output are mutually exclusive, retrying is limited to one attempt, and network errors should not be misread as credential errors. If _meta.json is missing or the skillCode is incorrect, the actual error should be checked manually rather than bypassing the authorization gate.