chore: expose canonical agent instructions to Gemini CLI

[devantler]

🤖 Generated by the Daily AI Engineer

Problem

go-template treats AGENTS.md as its canonical, cross-tool instruction file and already exposes it to Claude Code through the one-line CLAUDE.md shim. The template does not provide the equivalent GEMINI.md shim, so projects created from it do not give Gemini CLI the same canonical guidance. There is also no structural check preventing either shim from accumulating copied guidance and drifting from AGENTS.md.

This is a concrete child of the remaining template-suite parity theme in #104. The same inventory gap exists in dotnet-template; this issue keeps the first slice scoped to go-template so it can land independently.

Proposed direction

• Add root GEMINI.md as the exact one-line @AGENTS.md include.
• Add a lightweight template-repository validation that requires both CLAUDE.md and GEMINI.md to remain exact one-line shims.
• Document the shim inventory and validation command in AGENTS.md without duplicating canonical instructions into tool-specific files.

Acceptance criteria

• GEMINI.md exists at the repository root and contains exactly @AGENTS.md plus a trailing newline.
• CLAUDE.md remains the same exact shim.
• A failing-first structural test covers missing, copied, or otherwise non-canonical shim content and runs in the template-only scaffold validation workflow.
• AGENTS.md documents the two shims and the local validation command.
• Existing Go, scaffold, and workflow validation remains green.

Rough size

Small: one shim, one hermetic shell assertion, and the workflow/documentation wiring.

Part of #104.

[coderabbitai]

🔗 Related PRs

#146 - feat: scaffold OpenFeature Go SDK feature-flag support [merged]
#161 - ci: harden Copilot setup workflow [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!