feat(catalog-plugin): import music-catalog-diligence via subtree (3/3)

[sidneyswift]

The problem

The music-catalog-diligence plugin (royalty audits, rights diligence, valuation workpapers, IC memos for music catalog deals — 9 skills + 5 agents + 6 commands + scripts) lives in a separate GitHub repo. After PR #25, we have a unified marketplace — but the catalog plugin isn't in it. Customers can't install it from the new marketplace.

The fix

Imports the entire music-catalog-diligence plugin into plugins/ via git subtree (preserving full upstream history at commit a95f3dc). Registers it in marketplace.source.json so customers can install it.

After this lands:

/plugin install music-catalog-diligence@recoup

works in Claude Code, Codex, and Cursor.

How to review (mostly mechanical)

This PR has 2 commits, by design:

Commit 1 — the subtree import (~89 files)

• Single git-subtree merge bringing in the upstream plugin at commit a95f3dc.
• Verify the SHA matches. Don't read all 89 files.

Commit 2 — register + harden (~25 files)

• Adds the catalog plugin entry to marketplace.source.json, regenerates the 3 platform marketplace files.
• Plugin manifests across all 3 platforms get the right Recoup branding + the repository field pointed at recoupable/skills (was the pre-merge repo).
• Pulls in pass-1/pass-2 hardening from the original PR's bot reviews — Decimal monetary math, hooks deny-glob extension, validator tightening, doc fixes.

Files (113)

89 files come in mechanically via the subtree. Real review surface: ~25 files in commit 2.

What you should focus on

1. The subtree commit's SHA matches upstream a95f3dc.
2. The new plugin entry in marketplace.source.json looks right.
3. Commit 2's diff (~25 files) — that's the only meaningful surface.

Test plan

• python3 scripts/validate-manifests.py exits 0
• python3 plugins/music-catalog-diligence/scripts/test-helpers.py — 7 pass
• python3 plugins/music-catalog-diligence/scripts/test-validate-deal-workspace.py — 2 pass
• python3 plugins/music-catalog-diligence/scripts/test-normalize-royalty-statement.py — 7 pass
• Side-by-side screenshot: upstream recoupable/music-catalog-diligence@a95f3dc and the in-tree subtree at the same SHA
• Post-merge Loom: /plugin install music-catalog-diligence@recoup working in Claude Code

Merge order

PR 3 of 3. Stacked on #25 — auto-rebases when #25 lands. Save for last.

Merge preference

Please merge yourself when satisfied — I'd rather you push the button than approve.

🤖 Made with Cursor

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: ac64ebd0-558e-4817-a9fc-6e1ba757afbd

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/catalog-plugin

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[cubic-dev-ai]

6 issues found across 112 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="plugins/music-catalog-diligence/scripts/build-manual-review-queue.py">

<violation number="1" location="plugins/music-catalog-diligence/scripts/build-manual-review-queue.py:65">
P2: Filename fallback uses suffix matching (`endswith(name)`), which can create false positives for contributed files.</violation>
</file>

<file name="plugins/music-catalog-diligence/scripts/validate-normalized-ledger.py">

<violation number="1" location="plugins/music-catalog-diligence/scripts/validate-normalized-ledger.py:37">
P2: Avoid loading the entire ledger CSV into memory; iterate rows as a stream to prevent unnecessary memory growth on large files.</violation>
</file>

<file name="plugins/music-catalog-diligence/scripts/calculate-concentration.py">

<violation number="1" location="plugins/music-catalog-diligence/scripts/calculate-concentration.py:54">
P2: `top_n_pct` uses signed amounts after absolute-value ranking, which can misstate concentration when negative entries exist.</violation>
</file>

<file name="plugins/music-catalog-diligence/scripts/extract-pdf-statement.py">

<violation number="1" location="plugins/music-catalog-diligence/scripts/extract-pdf-statement.py:285">
P2: Path-based template detection is not OS-agnostic; Windows-style paths can prevent expected template matches.</violation>

<violation number="2" location="plugins/music-catalog-diligence/scripts/extract-pdf-statement.py:314">
P1: Amount parsing drops comma decimals, which can inflate extracted values for locale-formatted statements.</violation>
</file>

<file name="plugins/music-catalog-diligence/scripts/normalize-royalty-statement.py">

<violation number="1" location="plugins/music-catalog-diligence/scripts/normalize-royalty-statement.py:441">
P1: When no "Total" column exists and all component columns are also missing, `sum()` over an empty generator returns `0`, which then formats as `"0.00"` for `owner_net_amount`. This creates a false data point implying a computed zero-dollar amount in a row that has no financial data at all. Guard against the empty-components case so the output is `""` instead.</violation>
</file>

_{Note: This PR contains a large number of files. cubic only reviews up to 75 files per PR, so some files may not have been reviewed. cubic prioritizes the most important files to review.
On a pro plan you can use ultrareview for larger PRs.}

[cubic-dev-ai]

P1: Amount parsing drops comma decimals, which can inflate extracted values for locale-formatted statements.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At plugins/music-catalog-diligence/scripts/extract-pdf-statement.py, line 314:

<comment>Amount parsing drops comma decimals, which can inflate extracted values for locale-formatted statements.</comment>

<file context>
@@ -0,0 +1,607 @@
+            text = new_text
+            detected_currency = currency
+            break
+    text = text.replace(",", "").replace("$", "").replace("€", "").replace("£", "").replace("¥", "").strip()
+    if text.startswith("(") and text.endswith(")"):
+        text = "-" + text.strip("()")
</file context>

[cubic-dev-ai]

P1: When no "Total" column exists and all component columns are also missing, sum() over an empty generator returns 0, which then formats as "0.00" for owner_net_amount. This creates a false data point implying a computed zero-dollar amount in a row that has no financial data at all. Guard against the empty-components case so the output is "" instead.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At plugins/music-catalog-diligence/scripts/normalize-royalty-statement.py, line 441:

<comment>When no "Total" column exists and all component columns are also missing, `sum()` over an empty generator returns `0`, which then formats as `"0.00"` for `owner_net_amount`. This creates a false data point implying a computed zero-dollar amount in a row that has no financial data at all. Guard against the empty-components case so the output is `""` instead.</comment>

<file context>
@@ -0,0 +1,797 @@
+        "other": _to_float(first(row, "Other Income", "Other")),
+    }
+    total_value = _to_float(first(row, "Total"))
+    if total_value is None:
+        total_value = sum(v for v in components.values() if v is not None)
+    components_note = ", ".join(
</file context>

[cubic-dev-ai]

P2: Filename fallback uses suffix matching (endswith(name)), which can create false positives for contributed files.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At plugins/music-catalog-diligence/scripts/build-manual-review-queue.py, line 65:

<comment>Filename fallback uses suffix matching (`endswith(name)`), which can create false positives for contributed files.</comment>

<file context>
@@ -0,0 +1,218 @@
+        return True
+    # Try common alternate forms — bare filename, source/-prefixed.
+    name = str(manifest_entry.get("filename") or "")
+    if name and any(src.endswith("/" + name) or src.endswith(name) for src in ledger_sources):
+        return True
+    return False
</file context>

[cubic-dev-ai]

P2: Avoid loading the entire ledger CSV into memory; iterate rows as a stream to prevent unnecessary memory growth on large files.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At plugins/music-catalog-diligence/scripts/validate-normalized-ledger.py, line 37:

<comment>Avoid loading the entire ledger CSV into memory; iterate rows as a stream to prevent unnecessary memory growth on large files.</comment>

<file context>
@@ -0,0 +1,69 @@
+        reader = csv.DictReader(handle)
+        fieldnames = set(reader.fieldnames or [])
+        missing = sorted(REQUIRED_COLUMNS - fieldnames)
+        rows = list(reader)
+
+    errors: list[str] = []
</file context>

[cubic-dev-ai]

P2: top_n_pct uses signed amounts after absolute-value ranking, which can misstate concentration when negative entries exist.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At plugins/music-catalog-diligence/scripts/calculate-concentration.py, line 54:

<comment>`top_n_pct` uses signed amounts after absolute-value ranking, which can misstate concentration when negative entries exist.</comment>

<file context>
@@ -0,0 +1,219 @@
+def top_n_pct(ranked: list[tuple[str, float]], total: float, n: int) -> float | None:
+    if total == 0 or not ranked:
+        return None
+    head = sum(amount for _, amount in ranked[:n])
+    return round(head / total * 100, 2)
+
</file context>

Suggested change

	head = sum(amount for _, amount in ranked[:n])
	return round(head / total * 100, 2)
	head = sum(abs(amount) for _, amount in ranked[:n])
	denominator = sum(abs(amount) for _, amount in ranked)
	return round(head / denominator * 100, 2) if denominator else None

[cubic-dev-ai]

P2: Path-based template detection is not OS-agnostic; Windows-style paths can prevent expected template matches.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At plugins/music-catalog-diligence/scripts/extract-pdf-statement.py, line 285:

<comment>Path-based template detection is not OS-agnostic; Windows-style paths can prevent expected template matches.</comment>

<file context>
@@ -0,0 +1,607 @@
+
+def detect_template(rel_path: str, headers: list[str | None]) -> PDFTemplate | None:
+    header_text = _normalize_header(headers).lower()
+    rel_lower = rel_path.lower()
+    candidates: list[tuple[int, PDFTemplate]] = []
+    for template in TEMPLATES:
</file context>

Suggested change

	rel_lower = rel_path.lower()
	rel_lower = rel_path.replace("\\", "/").lower()