docs: point the upgrade guide's Copilot checks at .github/skills

[yxshee]

Description

The upgrade guide's troubleshooting step tells GitHub Copilot users to check ls -la .github/prompts/. On a default Copilot project that directory doesn't exist, so the documented check fails and suggests the commands were never installed.

Since #3976 (feat(copilot): default integration to skills), Copilot installs skills under .github/skills/ by default. .github/prompts/ (together with .github/agents/) only exists with the opt-in commands layout, --integration-options="--commands".

Reproduction with the current CLI:

$ specify init demo --integration copilot --ignore-agent-tools --non-interactive && cd demo
$ ls -la .github/prompts/
ls: .github/prompts/: No such file or directory
$ ls .github/
skills
$ ls .github/skills/
speckit-analyze
speckit-checklist
speckit-clarify
speckit-constitution
speckit-converge
speckit-implement
speckit-plan
speckit-specify
speckit-tasks
speckit-taskstoissues

With --integration-options="--commands", .github/ contains agents/ and prompts/ (speckit.analyze.prompt.md, …) instead.

Changes, all in docs/upgrade.md:

• the verification step now runs ls -la .github/skills/, with a comment naming .github/agents/ and .github/prompts/ for the --commands layout
• the three other places that list Copilot's command location now say .github/skills/

Heads-up: #4601 also edits the "permanently installed" sentence for a different reason (/speckit.x → /speckit-x), so whichever lands second has a one-line conflict.

Testing

• Tested locally with uv run specify --help

• Ran existing tests with uv sync && uv run pytest: docs-only change; no test reads docs/upgrade.md

• Tested with a sample project (if applicable): the specify init --integration copilot reproduction above, with and without --integration-options="--commands"

• npx markdownlint-cli2@0.23.2 docs/upgrade.md (the version pinned by DavidAnson/markdownlint-cli2-action v24.2.0 in lint.yml): 0 issues

AI Disclosure

• I did not use AI assistance for this contribution
• I did use AI assistance (fill in the disclosure below)

AI disclosure: Claude Code (model: Claude Opus 5.5, max reasoning effort, autonomous agent mode) opened this PR on behalf of @yxshee. A Claude Code subagent (same model, autonomous) found the drift by running the documented commands against the CLI and wrote the docs change. The lead agent reviewed the diff, ran markdownlint, and wrote this description. No human reviewed the change line by line before it was opened.

🤖 Generated with Claude Code