[allenporter]

Update the review skill since it seemed a little nitpicky on the last few rounds, and it's proposals made PRs worse.

Details:

• Clarify in .agents/skills/code-review/SKILL.md that device.features capability gating applies to option/command/trait exposure (and genuinely overloaded wire codes), rather than over-gating unambiguous read-only telemetry decoding (from_code_optional()).
• Add an explicit Adversarial Self-Check on Candidate Findings step before formatting or posting review feedback:
  • Verify suggested diffs against upstream/main baseline behavior and git log history so suggestions do not silently regress valid inputs or duplicate bounds.
  • Check sibling cases/codes in the same module to confirm stated invariants hold across the codebase.
  • Independently scrutinize claims from prior or competing PR reviews rather than carrying them forward uncritically.
  • Drop low-value docstring/comment wording nits and report cleanly when a PR has zero defects.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.