feat: add fable-5-playbook plugin

[kyle-sexton]

Closes melodic-software/medley#1293.

Publishes the fable-5-playbook plugin per docs/MIGRATION-PLAYBOOK.md (per-plugin gate + acceptance security review).

What ships

• plugins/fable-5-playbook/ — one knowledge skill (/fable-5-playbook:fable-5-playbook): Claude Fable 5's operating doctrine, authored by Fable 5 as introspected standing instructions. Core doctrine arms the session on invocation; twelve trigger-routed chapters under context/ (calibration, reasoning-moves, problem-framing, planning, execution, debugging, orchestration, verification, communication, recovery, context-economy, trust-and-authority) plus context/opus-adaptation.md for non-Fable models.
• Marketplace entry: category: learning, tags knowledge + component tags. Explicit version: 0.1.0 in plugin.json only.
• The source skill was already repo-agnostic (no repo-specific paths, tools, or slash references); the only content deltas from the source are the frontmatter description compressed to two sentences and one typo fix (mis-framed → misframed).

Gate evidence

• claude plugin validate ./plugins/fable-5-playbook --strict — PASS; claude plugin validate . --strict (catalog) — PASS
• claude plugin details token cost: ~207 tok always-on, ~4.7k on-invoke (single skill)
• markdownlint 0 errors (15 files); typos clean; editorconfig-checker clean
• --plugin-dir smoke test in a clean non-source repo — PASS: skill invoked, answered the four meta-rules (Precedence, One home per doctrine, Model adaptation, Silent application), namespace confirmed as fable-5-playbook:fable-5-playbook

Security review (playbook acceptance)

• Pure knowledge skill: no hooks, no MCP servers, no agents, no userConfig, no scripts, no state, zero network access.
• Cache isolation: all references are skill-internal (context/*.md self-citations); no ../ reach-outs.
• No PII / secrets.

Merge note: open PRs #53, #56, #57 also touch marketplace.json + root README.md — merge serially; later ones rebase/merge the two shared files (this branch is based on post-context7 main, so it conflicts only with those unmerged siblings).

---

Note

Low Risk
Markdown-only plugin packaging and catalog updates; no executable code, credentials, or runtime side effects beyond skill text loaded into Claude Code.

Overview
Adds the fable-5-playbook plugin to the marketplace and root catalog — a pure knowledge skill (no hooks, MCP, scripts, or userConfig) that arms sessions with Claude Fable 5’s operating doctrine.

SKILL.md loads core standing instructions on invoke (bare / full / chapter name), defines four meta-rules, effort floors, a chapter routing table, and cites twelve on-demand context/*.md chapters (calibration through trust-and-authority) plus mandatory opus-adaptation.md for non-Fable models. Plugin README.md and plugin.json (0.1.0, learning category) document install and scope.

^{Reviewed by Cursor Bugbot for commit 03fa989. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_1f826892-9eb2-471b-8680-bb8696fb89ce)

[claude]

Claude finished @kyle-sexton's task in 4m 0s —— View job

---

PR Review — feat/publish-fable-5-playbook

• Gather context and read changed files
• Review plugin.json manifest schema and metadata
• Review marketplace.json entry
• Review SKILL.md core doctrine content
• Review context/ chapter files for correctness and consistency
• Cross-check against CLAUDE.md design rules and playbook gate
• Security review validation
• Post final review findings

---

Overview

Solid PR. The content — 13 chapter files plus core doctrine — is high-quality, internally consistent, and follows the playbook's design rules for naming, scoping, and structure. The security posture is clean: pure knowledge skill, no hooks, no MCP servers, no userConfig, no scripts, zero network footprint. Gate evidence (validate, markdownlint, typos, smoke test) is thorough.

Two things need attention before merge, one substantive and one minor.

---

Finding 1 — Context file paths are not ${CLAUDE_PLUGIN_ROOT}-qualified (Medium)

File: plugins/fable-5-playbook/skills/fable-5-playbook/SKILL.md

The chapter routing table instructs the model to read files like context/calibration.md, context/problem-framing.md, etc. using bare relative paths. Meta-rule 3 also says read context/opus-adaptation.md NOW. These are model-level read instructions — when a trigger fires mid-session, the model must call the Read tool with a full path.

The CLAUDE.md design rule is explicit: "reference only files inside the plugin via ${CLAUDE_PLUGIN_ROOT}". The book-distill plugin (the closest structural analog) follows this correctly — it references its bundled template as ${CLAUDE_PLUGIN_ROOT}/skills/book-distill/templates/checklist.md. The migration playbook's cache-isolation caveat says the same: "Fix: bundle dependencies inside the plugin and reference them via ${CLAUDE_PLUGIN_ROOT}."

The concern: the smoke test confirmed the core doctrine loads and the four meta-rules answer correctly — but those answers are entirely within the SKILL.md body. The test did not exercise a mid-session trigger that lazy-loads a chapter (e.g., first code edit → execution.md). In an installed plugin at ~/.claude/plugins/cache/fable-5-playbook/…, the model may not be able to resolve context/calibration.md without the fully-qualified path.

Recommended fix — qualify all chapter references in the routing table and meta-rule 3:

`${CLAUDE_PLUGIN_ROOT}/skills/fable-5-playbook/context/problem-framing.md`

…and similarly for every entry in the routing table and for the context/opus-adaptation.md reference in meta-rule 3.

Fix this →

---

Finding 2 — New "learning" category not used elsewhere in the catalog (Low)

File: .claude-plugin/marketplace.json line 109

The entry uses "category": "learning", which is a new category not present in any other plugin. The most comparable plugin — book-distill — uses "category": "knowledge". Introducing a one-off category fragments discoverability; a user browsing by category won't find fable-5-playbook alongside book-distill.

Using "category": "knowledge" would be consistent with the catalog's existing taxonomy.

Fix this →

---

Minor notes (no action required)

• plugin.json description length. At ~226 characters it's notably longer than book-distill (133) and context7 (182). Functional, but may truncate in some UIs. Consider trimming the chapter list from the description and relying on the README.
• "fable-5" tag missing from marketplace entry. plugin.json keywords include "fable-5" but the marketplace tags array doesn't. Minor discoverability gap.
• opus-adaptation.md title vs. routing table. The file is titled "Running this playbook on Claude Opus 4.8" but the SKILL.md routing table says any non-Fable model should read it. The file's opening paragraph handles this gracefully ("If you are not Claude Opus 4.8: … apply the method"), so this is not a bug — just a slight mismatch between the title and the broader intent stated in the SKILL.md.

---

Security review (per playbook acceptance gate)

Surface	Verdict
Code execution (hooks/scripts)	✅ None
Remote MCP servers	✅ None
userConfig / credential handling	✅ None
Cache isolation	⚠️ Context file references need ${CLAUDE_PLUGIN_ROOT} (Finding 1)
Data egress / telemetry	✅ None
Provenance / third-party trust	✅ Author matches; MIT; no SaaS delegation

Finding 1 is the only open item before this can ship.

[cursor]

Stale comment

Risk: medium. Cursor Bugbot did not complete (check skipped due to usage limit), so automated review is incomplete. Human review is needed before merge.

  

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_935ed33c-4be3-4483-83ac-7e7ddf72d5c7)

[claude]

Claude finished @kyle-sexton's task in 1m 40s —— View job

---

PR Review — feat/publish-fable-5-playbook

• Gather context and read CLAUDE.md guidelines
• Read plugin.json manifest
• Read marketplace.json entry
• Read SKILL.md core doctrine
• Review context/ chapter files
• Cross-check design rules, security review, and prior findings
• Post final review

---

Overview

The plugin itself is high quality — pure knowledge skill with clean security posture, no hooks, no MCP servers, no userConfig, no state, zero network footprint. The content (core doctrine + 12 chapters + opus-adaptation) is internally consistent, well-structured, and follows the playbook's design rules for scoping and naming.

One blocking issue from the prior review remains unresolved. One finding from the prior review was incorrect and is retracted below.

---

Finding 1 — Context file paths not ${CLAUDE_PLUGIN_ROOT}-qualified (Blocking — still unresolved)

File: plugins/fable-5-playbook/skills/fable-5-playbook/SKILL.md

The routing table (lines 127–141) instructs the model to read chapter files as bare relative paths — context/calibration.md, context/problem-framing.md, etc. Meta-rule 3 (line 17) similarly says read context/opus-adaptation.md NOW.

CLAUDE.md's design rule is explicit: "reference only files inside the plugin via ${CLAUDE_PLUGIN_ROOT}". When installed at ~/.claude/plugins/cache/fable-5-playbook/…, the model must call Read with a fully-qualified path. A bare relative path will not resolve in that cache context.

The smoke test confirmed the core doctrine and four meta-rules answer correctly — but that is because those answers are entirely within the SKILL.md body itself. It did not exercise a mid-session trigger that lazy-loads a chapter, which is the path that actually fails.

Fix — qualify every context/*.md reference in the routing table and meta-rule 3 with ${CLAUDE_PLUGIN_ROOT}/skills/fable-5-playbook/:

`${CLAUDE_PLUGIN_ROOT}/skills/fable-5-playbook/context/problem-framing.md`

Apply to all 13 entries: the 12 routing table rows and the context/opus-adaptation.md reference in meta-rule 3.

Fix this →

---

Correction to prior review — Finding 2 retracted

The prior review flagged "category": "learning" as a new, inconsistent category. This was incorrect. Reading the current marketplace.json, both thariq-skills (line 108) and boris (line 113) already use "category": "learning". The fable-5-playbook entry is consistent with the existing catalog taxonomy. No change needed.

---

Minor notes (no action required)

• "fable-5" tag gap. plugin.json keywords include "fable-5" but the marketplace tags array omits it. Minor discoverability gap — worth adding to the marketplace entry alongside "doctrine".
• plugin.json description length. At ~226 characters, it's the longest description in the catalog. Functional, but may truncate in some UIs. The chapter list could move to the README.
• opus-adaptation.md title vs. intent. The file is titled "Running this playbook on Claude Opus 4.8" but the routing table and meta-rule 3 direct any non-Fable model to read it. The opening preamble handles this gracefully — not a bug, just worth noting for a future title revision.

---

Security review

Surface	Verdict
Code execution (hooks/scripts)	✅ None
Remote MCP servers	✅ None
userConfig / credential handling	✅ None
Cache isolation	⚠️ Context paths need ${CLAUDE_PLUGIN_ROOT} (Finding 1)
Data egress / telemetry	✅ None
Third-party trust / PII	✅ None

One open item: qualify the context file paths, then this is ready to ship.

---

[cursor]

Risk: medium. Cursor Bugbot did not complete successfully (check skipped), so automated review is incomplete. Human review is needed before merge; no reviewers were assigned because no suitable non-author reviewers are available in this repository.

  

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 03fa9893a7

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]

Route durable notes outside consumer repos

When this chapter fires for a long session or a conclusion that took more than ~5 tool calls, the skill makes writing a “durable work note” mandatory, but the plugin has no configured/documented storage path and its README describes it as having no state. In a marketplace install, a model following this instruction in an arbitrary consumer project has to choose where to create or update that note, which can leave unexpected untracked or committed artifacts in the user’s repo even though they only loaded a knowledge skill. Please make the note write ask-first/configured, or route it to a documented plugin data location.

Useful? React with 👍 / 👎.