copilot-surface-auditlisted

# Skill: Copilot Surface Audit A Copilot model recommendation is only as good as its surface scope. "Model X is available in Copilot" is an incomplete claim until the surface (completions vs. chat vs. coding agent vs. cloud agent vs. mobile) and the plan (Free/Pro/Business/Enterprise) are both confirmed. This skill ensures every recommendation is surface-scoped and plan-gated. ## Step 1 — Identify the active surface(s) Ask or infer from context which Copilot surface(s) the developer actually uses: | Surface | Entry point | Model picker location | |---|---|---| | Completions | IDE inline ghost-text | No picker — org policy or default applies | | Chat IDE | IDE sidebar / slash commands | Model picker in chat header | | Coding agent | IDE-initiated autonomous run | Separate agent surface — plan-gated | | Cloud agent | GitHub.com / Copilot Workspace | Separate cloud picker — may differ from IDE | | Mobile | GitHub Mobile app | Mobile-specific subset — narrowest picker | **Do not conflate surfaces.** A model available in chat IDE may not be available for the coding agent; cloud agent has its own availability list. ## Step 2 — Confirm plan tier | Plan | Key capabilities | Notes | |---|---|---| | Free | Subset of models; monthly usage cap | Narrowest model picker | | Pro | Full IDE model picker; no monthly cap | Baseline for professional use | | Business | Org policy controls; team seat management | Adds model rules + audit log | | Enterprise | Full controls; org model allow/