feat(output): add --suppress-reads flag

[osolmaz]

Supersedes #136 to land the same change from an origin-hosted branch after the contributor fork branch stopped reporting checks reliably after maintainer updates.

This replacement keeps the contributor's feature commit, adds the required changelog entry, and preserves credit to @hayatosc.

Original contribution:

• feature commit: cc5883a by @hayatosc
• changelog note: f6daad2 by @osolmaz

Behavior is unchanged relative to the contributor branch: this still adds --suppress-reads so text and JSON output can mask raw file-read bodies while leaving normal tool activity visible.

Follow-up after merge:

• close #136 as superseded by this PR, with credit to @hayatosc.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f6daad2bf7

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Prefer explicit non-read kind over title inference

isReadLikeTool currently treats a tool as read-like whenever the title contains read-ish substrings, even if kind is explicitly set to a non-read value. With --suppress-reads, this causes non-read tools to have their output masked (e.g., { kind: "edit", title: "Edit README.md" } matches read in the title and gets suppressed), which hides normal tool activity unexpectedly in both text and JSON output. The explicit kind should take precedence over heuristic title matching.

Useful? React with 👍 / 👎.