v0.5.0 — model-facing strip note across all agents

What's Changed

silence now tells the coding agent why it stripped your comments, so the model stops re-adding them on the next edit.

After a post-edit strip, the hook emits a short model-facing note — "silence stripped N comments. Don't re-add. Prefer self-explanatory code." — through each agent's native channel:

• Claude Code and Codex read it from the hook's stdout (hookSpecificOutput.additionalContext)
• Opencode and Pi plugins splice it into the tool result the model sees

It fires only when comments were actually removed and stays silent in your terminal — the model gets the nudge, you don't get the noise.

PRs

• Feed the agent an additionalContext note after stripping by @hasparus in #13
• Extend the note to Codex, Opencode, and Pi by @hasparus in #14

Full Changelog: v0.4.0...v0.5.0