docs(security): add security and sandboxing guidance

[amaad31]

Adds a dedicated security/sandboxing page (#1120).

Used the structure @ suggested in the issue: separates what Prime Agent enforces from what the operator needs to isolate, with concrete guidance (container/VM, don't mount $HOME/SSH keys/Docker socket, non-root user) and a pre-run checklist.

Two things worth flagging for review:

• No built-in command approval/allowlist mechanism exists today, so the doc says that explicitly instead of implying one.
• Found --no-session while checking this, useful for disposable one-shot runs, added it under Workspace Boundaries and Persisted State.

Linked from docs/index.md and the README.

Fixes #1120

Note

Add security and sandboxing guidance to coding agent docs

• Adds a new security.md covering the trust model, recommended container/VM isolation, workspace boundaries, tool restrictions, persisted state locations, and a pre-run checklist.
• Updates packages/coding-agent/docs/index.md to include a 'Security and sandboxing' link in the Start Here section.
• Adds a warning link in README.md pointing to the new security doc for isolation recommendations.

^{Macroscope summarized ebc4405.}

[sethkarten]

Thank you for the report and proposed work. This root cause is now covered by maintainer-owned stacked PR #1159, authored independently from upstream/main.

We did not inspect or reuse this PR's diff, branch, commits, implementation code, or tests; its public description/comments were used only as a bug report. To keep one review surface, this PR is superseded by #1159 and is being closed.

The complete review stack is #1158–#1165. It is being left unmerged for human review after CI and review-bot findings are cleared.