docs(plugins): catalog official and community plugins and compare with ai-config (closes #99)

[dem-extra1]

Issue #99 asks for a catalog of useful plugins: start with the official Claude plugins, branch out to community marketplaces, and compare and contrast with the lab's in-house ai-config. A follow-up comment asks what the ralph-loop plugin is and whether it is like ai-config's ardi.

Why: the existing useful-plugins section lists four specific tools (SemBr, ponytail, Contextify, ai-config) but gives readers no map of the official marketplace or of the community ecosystem, and no basis for deciding what to install alongside ai-config.

This PR extends the useful-plugins section with:

• a catalog of the official Anthropic plugin marketplace (name, purpose, one-line verdict)
• notable community marketplaces and plugins
• a compare-and-contrast subsection against Morrison-Lab/ai-config
• an explicit answer to the ralph-loop question

Closes #99

🤖 Generated with Claude Code

[dem-extra1]

Pushed with the pre-push guard's ALLOW_UNREVIEWED_PUSH=1 override. The adversarial-reviewer subagent the guard expects is not registered in this session, so the adversarial pre-push review was performed by the Google Antigravity CLI (agy --print) against the full branch diff; real findings were fixed before this push. The @claude review workflow and a requested Copilot review are the independent reviews of record.

Posted by Claude Code (AI agent) --- not written by a human.

[github-actions]

👀 Picked up by workflow run #34437190287. Setup runs first; Claude itself responds after that.

[Copilot]

🟡 Changes recommended

Several new bibliography entries and parts of the new catalog content conflict with established repo conventions and contain a concrete internal inconsistency that should be corrected before merge.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

This PR expands the “Useful plugins” documentation by introducing a dedicated plugin-catalog section that maps Claude Code’s official/community plugin ecosystems and contrasts them with the lab’s in-house ai-config, including an explicit explanation of ralph-loop vs ardi.

Changes:

• Adds a new “plugin catalog” section to the Coding Agents chapter and links it from the existing useful-plugins material.
• Introduces a new plugin-catalog.qmd page fragment covering Anthropic marketplaces, notable community plugins, and an ai-config comparison (including ralph-loop).
• Extends the bibliography and spellcheck wordlist to support the new content.

File summaries

File	Description
chapters/coding-agents.qmd	Adds a new top-level section that includes the plugin catalog fragment.
chapters/ai-tools/useful-plugins.qmd	Adds a cross-reference pointing readers to the new catalog section.
chapters/ai-tools/plugin-catalog.qmd	New catalog content: marketplaces overview, community notes, ai-config comparison, and ralph-loop explanation.
references.bib	Adds new citations for marketplaces/plugins/blog post and ai-config.
inst/WORDLIST	Adds newly introduced proper nouns/terms to satisfy spelling checks.

Review details

Suppressed comments (4)

references.bib:275

• Bibliography convention: GitHub repository citations in references.bib use howpublished = {Software} (not {GitHub repository}), so this entry should match for consistency.

  howpublished = {GitHub repository},

references.bib:284

• Bibliography convention: GitHub repository citations in references.bib use howpublished = {Software} (not {GitHub repository}), so this entry should match for consistency.

  howpublished = {GitHub repository},

references.bib:320

• Bibliography convention: GitHub repository citations in references.bib use howpublished = {Software} (not {GitHub repository}), so this entry should match for consistency.

  howpublished = {GitHub repository},

references.bib:329

• Bibliography convention: GitHub repository citations in references.bib use howpublished = {Software} (not {GitHub repository}), so this entry should match for consistency.

  howpublished = {GitHub repository},

• Files reviewed: 5/5 changed files
• Comments generated: 6
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

🟢 Approval recommended

The changes are documentation-focused and internally consistent; only a minor punctuation/style nit was identified.

Review details

Suppressed comments (1)

chapters/ai-tools/useful-plugins.qmd:72

• The new cross-reference sentence reads like a run-on and is missing closing punctuation. Consider merging it into a single sentence with a comma and an ending period for readability and consistent style.

@sec-ai-plugin-catalog maps the official and community marketplaces these four sit in
and sets each against `ai-config`.

• Files reviewed: 5/5 changed files
• Comments generated: 0 new
• Review effort level: Lite

[Copilot]

🟡 Changes recommended

references.bib contains malformed/incomplete BibTeX entries that will break bibliography parsing and downstream renders.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Review details

Suppressed comments (1)

references.bib:313

• ai_config_readme in references.bib is missing its closing } (and a note field), so the entry currently runs into the following @misc{gsd_core,...} and will break BibTeX parsing.

@misc{ai_config_readme,
  title = {ai-config: portable AI agent config},
  author = {{Morrison Lab}},
  year = {2026},
  howpublished = {Software},
  url = {https://github.com/Morrison-Lab/ai-config},
@misc{gsd_core,

• Files reviewed: 5/5 changed files
• Comments generated: 1
• Review effort level: Lite

[Copilot]

🔵 Needs a closer look

The new Quarto prose introduces several long comma-separated technical lists that conflict with the repo’s documented list-formatting conventions, and should be reformatted for consistency and readability.

Review details

Suppressed comments (4)

Previously missed (1) — in code that hasn't changed since the last review.

chapters/ai-tools/plugin-catalog.qmd:44

• This parenthetical list has 3+ technical items; repository style guidance prefers a bullet list (with a blank line before it) for readability in .qmd prose (see .github/copilot-instructions.md:24-51).

This issue also appears in the following locations of the same file:

• line 84
• line 158

chapters/ai-tools/plugin-catalog.qmd:88

• This inline list of 12 *-lsp plugin names is long enough that it’s hard to scan in rendered output. The repo’s style guidance recommends bullet lists for 3+ technical items in .qmd prose (see .github/copilot-instructions.md:24-51).

Twelve `*-lsp` plugins
(`clangd`, `csharp`, `gopls`, `jdtls`, `kotlin`, `lua`, `php`, `pyright`, `ruby`, `rust-analyzer`, `swift`, `typescript`)
connect a Language Server Protocol server so that Claude sees type errors and missing imports after every edit
and can jump to definitions instead of grepping
[@claude_code_discover_plugins].

chapters/ai-tools/plugin-catalog.qmd:162

• This is a long comma-separated list of MCP server names (3+ technical items). Per repo prose style guidance, this should be a bullet list (with a blank line before it) to improve readability in the rendered Quarto output (see .github/copilot-instructions.md:24-51).

The `external_plugins/` directory and many third-party entries bundle a pre-configured MCP server:
`github`, `gitlab`, `atlassian`, `asana`, `linear`, `notion`, `figma`, `slack`, `sentry`,
`vercel`, `firebase`, `supabase`, `context7`, `playwright`, `serena`, `terraform`,
and messaging bridges for Discord, Telegram, and iMessage
[@claude_code_discover_plugins].

chapters/ai-tools/useful-plugins.qmd:72

• This sentence ends without closing punctuation, which reads like an unfinished thought in the rendered prose. Add a trailing period (and keep semantic line breaks intact).

@sec-ai-plugin-catalog maps the official and community marketplaces these four sit in
and sets each against `ai-config`.

• Files reviewed: 5/5 changed files
• Comments generated: 0 new
• Review effort level: Lite

[github-actions]

Claude finished review — View run

---

Review: PR #217 — docs(plugins): catalog official and community plugins and compare with ai-config

Scope of this round. Per the computed range 2f8fca0a..d45cebbd, this round's commits (b432acc merge of origin/main, plus the pass-through of already-merged PRs #211 and #209) bring chapters/ai-tools/opencode-ecosystem.qmd and chapters/ai-tools/posit-dev-ai-repos.qmd into this branch's history. I verified both files are byte-identical between origin/main and this branch's head (git diff origin/main d45cebb -- <those two files> → 0 lines) — they arrived purely through a routine merge-from-main sync, not new authorship on this branch, and aren't part of the PR's actual diff against main (gh/saved diff confirms neither file appears in it). No new review needed there; they were already reviewed on their own PRs.

The previously-flagged references.bib corruption is fixed. The prior round (2026-09-10T08:17:22Z, comment) found the obra_superpowers and ai_config_readme entries truncated mid-block with missing fields and closing braces, breaking BibTeX parsing. I re-checked the current file:

• Both entries are now complete and well-formed (references.bib:713-720, references.bib:722-729), each with a proper closing }.
• File-wide brace count is balanced: 624 { vs. 624 }.
• Ran an actual BibTeX parser (bibtexparser 2.0.0) against the full file: 81 entries parsed, 0 failed blocks.

That verdict item is resolved and should not be re-raised.

chapters/ai-tools/plugin-catalog.qmd is unchanged content. git diff 46024ad d45cebb -- chapters/ai-tools/plugin-catalog.qmd is empty — this file is byte-identical to the commit already fact-checked in the two earlier rounds (superpowers skill list, marketplace categories, all citation keys, cross-refs). No re-verification needed; that fact-check still stands.

New, non-blocking finding: duplicate bibliography entry for the same source.

references.bib now contains two separate @misc entries pointing at the exact same URL (https://github.com/obra/superpowers), with near-identical content:

• references.bib:280-287 — key superpowers (title "Superpowers: An agentic skills framework and software development methodology"), added by unrelated, already-merged content in chapters/ai-tools/custom-harnesses.qmd (cited there as [@superpowers], e.g. chapters/ai-tools/custom-harnesses.qmd:38,108).
• references.bib:713-720 — key obra_superpowers (title "Superpowers"), added by this PR and cited in its own chapters/ai-tools/plugin-catalog.qmd:186.

This is not a parse error and doesn't break the build — both keys are syntactically valid and distinct, so citeproc will render both fine. It's a content-hygiene issue: two authors independently added a reference for the same GitHub repo under different keys and slightly different titles, and the two only now coexist after this branch synced with main. Worth consolidating onto one key (probably superpowers, since it was in main first) before merge, but it's a minor cleanup, not a blocker — flagging as non-blocking per the instruction to fix at the source rather than let it linger, since a future citation of "Superpowers" elsewhere in the site risks picking whichever key happens to exist.

Everything else already verified in prior rounds is unchanged and still holds: bullet-list formatting (re-verified — every - list in plugin-catalog.qmd has a blank line above it), balanced code fences (one open/close pair), all citation keys used in the new chapter resolve to references.bib entries, all @sec-* cross-references resolve to defined labels, and the new inst/WORDLIST entries (Huntley, Wiggum, backpressure, composable, greenfield, preflight) are each actually used in the prose and the file remains sorted with no duplicates.

Verdict

Ready for merge

The previously-blocking references.bib corruption (missing closing braces on obra_superpowers/ai_config_readme) is confirmed fixed. The one new item found — a duplicate superpowers/obra_superpowers bib entry for the same source — is a minor, non-blocking cleanup rather than a defect that breaks rendering or violates a hard CLAUDE.md rule.

Structured Review Data (JSON)

{
  "schema_version": "1.1",
  "reviewer": "claude",
  "commit_sha": "d45cebbd022dd871df2d6d8a55aca90bf973f006",
  "verdict": "CLEAN",
  "findings": [
    {"file": "references.bib", "line": 713, "category": "doc", "message": "obra_superpowers duplicates the pre-existing superpowers entry (references.bib:280) — both cite https://github.com/obra/superpowers under different keys and slightly different titles, one from this PR's plugin-catalog.qmd and one from already-merged custom-harnesses.qmd content. Non-blocking: valid BibTeX, renders fine, but worth consolidating onto one key."}
  ],
  "detailed_assessment": "references.bib now parses cleanly with bibtexparser (81 entries, 0 failed blocks, 624/624 balanced braces), confirming the prior round's flagged missing-closing-brace corruption on obra_superpowers and ai_config_readme is fixed; the only remaining defect is a harmless duplicate citation entry for the same obra/superpowers URL under two different keys.",
  "holistic_assessment": "opencode-ecosystem.qmd and posit-dev-ai-repos.qmd entered this branch purely via a routine merge from main (already reviewed on PRs #211/#209, byte-identical to origin/main) and are out of this PR's actual diff scope, while plugin-catalog.qmd itself is byte-identical to the commit already fact-checked in two prior review rounds, so no new prose or citation validation was required beyond the bib-integrity regression check."
}

Reviewed commit: b432acc

[github-actions]

💰 Cost: $0.6331 (review) — run

[Copilot]

🟡 Changes recommended

There’s at least one documentation/bibliography correctness/robustness issue (a raw GitHub content URL citation) that should be fixed before merging to reduce link fragility and potential CI link-check failures.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Review details

• Files reviewed: 5/5 changed files
• Comments generated: 1
• Review effort level: Lite

[github-actions]

PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-09-10 09:27 UTC