fix(claude-memory): count newline-named topic files once in scope-report

[kyle-sexton]

Closes #3444

Summary

scope-report.sh counted topic files with find | wc -l, which counts newlines in the stream, so a filename containing an embedded newline was counted twice.

Fix

Replaced the pipeline with the sibling null-delimited idiom: find -print0 plus read -r -d ''. Plugin bumped to 0.11.12.

Verification

Added a fixture whose name contains a literal newline and asserted topic files: 2. Confirmed the new assertion fails against the pre-fix script and passes after. scripts/affected-tests.sh --run selected the co-located suite; all 22 checks passed.

Related

N/A

Note: this PR and #3374 both touch scope-report.test.sh. Merge #3374 first, then rebase this one.

  

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-02T06:19:06.050729Z	856cbe4	PR opened

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[claude]

Claude finished @kyle-sexton's task in 54s —— View job

---

Security review complete

• Gather context (PR diff, custom instructions)
• Run skip-gate check
• Review changed files for security-relevant surface
• Post findings

No security issues found.

This PR only touches plugins/claude-memory/skills/stateless/scripts/scope-report.sh (counting logic), its test file, plugin.json (version bump), and CHANGELOG.md. The fix replaces find ... | wc -l with the null-delimited find -print0 / while IFS= read -r -d '' idiom already used elsewhere in this plugin (scope-report.sh#L114-L118).

Notes from review, none rising to a reportable finding:

• No new user-controllable input reaches this script; mem_dir comes from the plugin's existing resolver, unchanged by this diff.
• The null-delimited read loop is the correct fix for embedded-newline/whitespace filenames and introduces no quoting or injection issues (no eval, no unquoted expansion of filenames into a command).
• No changes to permissions, trust boundaries, or GitHub Actions workflows.

Skip-gate condition 2 (no security-relevant surface after reading the diff) applies here.

[github-actions]

Last security-reviewed head: 856cbe4d43af42ac4170f5f658e28f22d73bc31f. On the next push, the relevance gate compares only the commits since this SHA; delete this comment to force a full re-review.

[claude]

Claude finished @kyle-sexton's task in 1m 7s —— View job

---

Code review complete

• Gather context (PR diff, changed files)
• Understand the request (invoke /review:code-review per CI lane instructions)
• Review changed files for correctness/maintainability
• Post findings

Scope: plugins/claude-memory/skills/stateless/scripts/scope-report.sh, its test file, plugin.json, CHANGELOG.md.

Summary: The fix replaces find | wc -l (which counts newlines in the stream, not files) with the null-delimited find -print0 + while IFS= read -r -d '' idiom already used by the sibling script enumerate-all-projects.sh (scope-report.sh#L114-L118), so a topic filename containing an embedded newline now counts once instead of twice. This exactly mirrors the existing idiom at enumerate-all-projects.sh:56-60, keeping the two scripts consistent.

The regression test (scope-report.test.sh Case 4b) builds on the existing Case 4 fixture, adds a newline-named .md file, asserts topic files: 2, and cleans up after itself — a solid, minimal repro of the bug. plugin.json and CHANGELOG.md updates are consistent with each other and with the changelog's existing formatting conventions.

I read the full diff and the surrounding script/test files. I was not able to execute the test suite myself in this sandboxed environment (bash execution required approval that wasn't available here), so the PR's "22 checks passed" claim is author-reported and unverified by me — the logic reads correct on inspection, and matches an idiom already proven out in the sibling script.

No high-signal findings. The change is small, targeted, matches the high-signal bar's exclusions (no style nits, no pre-existing-issue scope creep), and correctly limits itself to the one broken counting path.

One non-blocking note already disclosed in the PR body: this PR and #3374 both touch scope-report.test.sh, so merge order matters as the author says.

[github-actions]

Claude has reviewed this PR 1 time. The lane skips further automatic reviews after 5; deleting this comment resets the count.