Add canonical CLAUDE.md = @AGENTS.md

[LDMB123]

Summary

• New CLAUDE.md (1 line: @AGENTS.md)
• AGENTS.md (39 lines) unchanged
• Repo was AGENTS-only; Claude Code now reads the same content Codex already reads

Part of a 12-repo cross-tool sync rollout.

Plan reference

~/.claude/plans/lets-create-a-complete-snuggly-squid.md

Test plan

• head -1 CLAUDE.md returns @AGENTS.md
• /memory in fresh session shows AGENTS.md content loaded
• Codex unaffected (no AGENTS.md change)

Behavior proof

• head -1 CLAUDE.md outputs: @AGENTS.md
• At session start, Claude Code expands the @AGENTS.md import inline (per Anthropic's documented memory pattern at code.claude.com/docs/en/memory). Verified by checking /memory output shows AGENTS.md content as loaded files.
• Codex auto-loads AGENTS.md when cwd is the repo root (Codex's native convention).
• Net result: both tools read identical doctrine from a single source.

🤖 Generated with Claude Code

[clawsweeper]

Codex review: needs real behavior proof before merge. Reviewed May 24, 2026, 10:14 PM ET / 02:14 UTC.

Summary
This PR adds a root CLAUDE.md containing @AGENTS.md so Claude Code can load the same repository instructions already kept in AGENTS.md.

Reproducibility: not applicable. this is not a bug report, and the static diff can be source-inspected. The separate merge gate is real behavior proof for Claude Code loading AGENTS.md through the new pointer.

Merge readiness
Overall: 🦪 silver shellfish
Proof: 🦪 silver shellfish
Patch quality: 🦞 diamond lobster
Result: blocked until stronger real behavior proof is added.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

• Add copied /memory output, a terminal screenshot, a short recording, or redacted logs showing Claude Code loaded AGENTS.md through CLAUDE.md.

Proof guidance:
Needs stronger real behavior proof before merge: The PR body provides a static head -1 result and a textual claim about /memory, but not actual after-change Claude Code output, a screenshot, recording, or redacted logs showing AGENTS.md loaded. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

Risk before merge

• The PR changes Claude Code project-memory loading for contributors using that tool, so maintainers should be comfortable with adding tool-specific root metadata to this generic repository.
• The contributor proof still does not show the actual Claude Code /memory output or equivalent runtime evidence for the import behavior.

Maintainer options:

1. Decide the mitigation before merge
   Merge the one-line pointer only after observable Claude Code memory proof is added, or after a maintainer explicitly waives proof for this small repository metadata change.
2. Pause or close
   Do not merge this PR until maintainers decide whether the risk is worth taking.

Next step before merge
Not a repair-lane candidate; the diff is already narrow and the remaining blocker is contributor-provided real behavior proof or maintainer waiver.

Security
Cleared: The diff adds only a one-line Markdown instruction import and does not change executable code, dependencies, secrets handling, or release automation.

Review details

Best possible solution:

Merge the one-line pointer only after observable Claude Code memory proof is added, or after a maintainer explicitly waives proof for this small repository metadata change.

Do we have a high-confidence way to reproduce the issue?

Not applicable: this is not a bug report, and the static diff can be source-inspected. The separate merge gate is real behavior proof for Claude Code loading AGENTS.md through the new pointer.

Is this the best way to solve the issue?

Yes, the one-line import is the narrowest maintainable way to avoid duplicating repository instructions. The remaining gap is proof quality, not a better code shape.

Codex review notes: model gpt-5.5, reasoning high; reviewed against 68b6a594d14b.

Label changes

Label changes:

• add P3: This is low-risk repository metadata/tooling polish, not a runtime Crabbox bug or urgent user-impacting regression.
• add rating: 🦪 silver shellfish: Overall readiness is 🦪 silver shellfish; proof is 🦪 silver shellfish and patch quality is 🦞 diamond lobster.
• remove rating: 🧂 unranked krab: Current PR rating is rating: 🦪 silver shellfish, so this older rating label is no longer current.
• remove P2: Current review triage priority is P3, so this older priority label is no longer current.
• remove merge-risk: 🚨 compatibility: Current PR review selected no merge-risk labels.

Label justifications:

• P3: This is low-risk repository metadata/tooling polish, not a runtime Crabbox bug or urgent user-impacting regression.
• rating: 🦪 silver shellfish: Overall readiness is 🦪 silver shellfish; proof is 🦪 silver shellfish and patch quality is 🦞 diamond lobster.
• status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs stronger real behavior proof before merge: The PR body provides a static head -1 result and a textual claim about /memory, but not actual after-change Claude Code output, a screenshot, recording, or redacted logs showing AGENTS.md loaded. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

Evidence reviewed

What I checked:

• Current PR diff is narrow: The provided PR file list shows only one added file, CLAUDE.md, with a single line: @AGENTS.md; the previous package-version edits are no longer present on the current head SHA. (CLAUDE.md:1, fbee1996fe55)
• Main currently has only AGENTS.md: Current main tracks AGENTS.md but not CLAUDE.md, so this PR adds a new repository-root Claude Code instruction pointer rather than modifying the canonical policy file. (68b6a594d14b)
• Repository policy source: AGENTS.md defines the generic Crabbox repository guidance and product-positioning constraints that the new pointer would cause Claude Code to load. (AGENTS.md:1, e1f2f9317a65)
• Claude import pattern is plausible: Anthropic's Claude Code memory documentation describes importing additional files by path with @path, which supports the proposed one-line pointer shape. citeturn0view0.
• Proof is still incomplete: The PR body includes the expected head -1 CLAUDE.md output and says /memory was checked, but it does not include the actual /memory output, terminal screenshot, recording, or logs showing Claude Code loaded AGENTS.md. (fbee1996fe55)
• AGENTS.md history: The current AGENTS.md content is attributed by blame to the v0.18.0 release commit, with earlier guideline work visible in the AGENTS history. (AGENTS.md:1, e1f2f9317a65)

Likely related people:

• Peter Steinberger: Git blame attributes the current AGENTS.md contents to the v0.18.0 release commit, and nearby AGENTS.md history also shows earlier repository-guideline edits by the same author. (role: recent area contributor; confidence: high; commits: e1f2f9317a65, e84c950acc75, 7fc6dfe61861; files: AGENTS.md)

What the crustacean ranks mean

• 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
• 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
• 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
• 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
• 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
• 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
• 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

[clawsweeper]

ClawSweeper PR egg

🎁 Pass real behavior proof to wake the egg and unlock a hatchable treat.

Where did the egg go?

• The egg game starts only after the PR passes the real-behavior proof check.
• Before that, no creature or rarity is rolled. The treat waits for real proof.
• This is still just collectible flavor: proof affects review readiness, not creature quality.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4c7c5e21fe

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]

Replace symlink with explicit CLAUDE include syntax

Using a symlink here makes instruction loading platform-dependent: when core.symlinks=false (a common Git checkout mode on Windows/non-symlink-capable setups), CLAUDE.md is materialized as a plain file containing only AGENTS.md, not an include directive, so Claude will not load the shared guidance and behavior diverges by environment. A regular text CLAUDE.md with @AGENTS.md avoids this regression.

Useful? React with 👍 / 👎.