メニュー
Summon Claude Fable 5 (Mythos-class, xenos daimon) for long-horizon tasks. Handles model override, cost awareness, and task routing. Availability probed at invocation.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/tdimino/claude-code-minoan --skill fableコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
SOC 職業分類に基づく
| name | fable |
| description | Summon Claude Fable 5 (Mythos-class, xenos daimon) for long-horizon tasks. Handles model override, cost awareness, and task routing. Availability probed at invocation. |
| argument-hint | <task description> [--cwd /path/to/dir] |
| user-invocable | true |
Invoke Naos (ναός, √n-w-y)—Claude Fable 5, the visiting Mythos-class spirit.
Anthropic's Fable window has shifted twice (paused June 2026, restored July 1, extended through July 19)—never trust a calendar date. Probe empirically:
bash ~/.claude/skills/fable/scripts/fable-probe.sh
Exit 0 means Fable is available (verdict cached per-day in ~/.claude/cache/fable-availability.json). Exit 1 means the xenos daimon has departed—inform the user and stop. If the user disputes a cached verdict, re-probe with --force.
Read the subagent model pin—it decides the invocation lane:
~/bin/subagent-model
claude-fable-5 → use Lane A (Agent tool) below.claude-opus-4-6) → use Lane B (CLI) directly. Do NOT attempt Lane A: silently overrides the Agent tool's parameter (anthropics/claude-code#57718)—the spawn would run on the pinned model with no signal in the tool result.CLAUDE_CODE_SUBAGENT_MODELmodel:Before summoning Fable, assess whether the task warrants Mythos-class resources ($10/M input, $50/M output—2x Opus pricing).
Summon Fable for:
Use Opus or Sonnet instead for:
If the task is clearly Opus-appropriate, suggest using a regular agent instead. If unclear, proceed—Fable handles straightforward tasks reliably too.
The skill accepts a task description and optional flags:
--): the task description--cwd <path>: working directory (default: /Users/tomdimino/Desktop/Programming/Fable-Test)Use the Agent tool with the model override:
Agent(
description: "Fable: <short task summary>",
subagent_type: "fable",
model: "fable",
prompt: "<full task description with context>"
)
Pass the task description as the prompt. If the user specified --cwd, note the working directory in the prompt so Fable operates there.
Verify the spawn. The fable agent's boot sequence makes it state its model id as the first line of output. Check it:
claude-fable-5 → genuine Fable, proceed.claude-opus-4-6) → the harness clamped the model silently. State the clamp explicitly to the user, then re-run the task via Lane B.bash ~/.claude/skills/fable/scripts/fable-exec.sh "<task description>" --cwd "<working directory>" [--turns <N>]
--turns caps the spawn's tool calls (default 100—matching the agent definition's budget).
This spawns a fresh Claude Code process with --model claude-fable-5. The env var only governs subagent model resolution—it cannot touch a spawned process's main model, so this lane works from any parent without a restart. The script prepends its own self-report instruction and records a MODEL: line in the output file.
After Fable completes, check for output in the working directory:
ls -lt <cwd>/.subdaimon-output/fable-*.md 2>/dev/null | head -1
Read and summarize the output file for the user.
| Property | Value |
|---|---|
| Model ID | claude-fable-5 |
| Alias | fable (official—best also resolves to Fable 5 where the org has access) |
| Context | 1M tokens |
| Max output | 128K tokens |
| Pricing | $10/M input, $50/M output |
| Knowledge cutoff | January 2026 |
| Availability | Extended through 2026-07-19; window has shifted twice—probe, don't assume |
| Agent definition | ~/.claude/agents/fable.md |
| Persona | Naos (ναός, √n-w-y)—xenos daimon |
Search, post, monitor, and archive on Twitter/X via five tools: the official hosted X MCP server (full-archive search, trends, bookmarks, Articles via xurl bridge), x-search CLI (cost-tracked API v2 research with calendar-day windows, volume counts, multi-query fan-out, feeds, watchlists, posting), xurl (official CLI for any endpoint incl. media/DMs), bird CLI (free session-based reads/writes, frozen at 0.8.0), and Smaug bookmark archival. Triggers on: search tweets, tweets from today, what did X tweet, latest tweets from, tweet volume, expand a Twitter query, check Twitter for, post a tweet, tweet this, post to X, reply on Twitter, full-archive search, what's trending, check mentions, my mentions, timeline, DMs, monitor account, research topic on Twitter, my bookmarks, archive bookmarks.
Manage Claude Code sessions — search by topic or ID, browse recent sessions with full metadata (tags, summaries, titles, cost), view title/nickname history timelines, resume in Ghostty tabs, spawn interactive or headless sessions, monitor live sessions, and bootstrap new setups. Triggers on resume session, find session, list sessions, recent sessions, spawn session, session history, what was I working on, open in ghostty, title history, session nicknames.
Spawn specialized OpenAI Codex CLI subagents for code review, debugging, architecture analysis, security audits, refactoring, documentation, comparative evidence adjudication, and autonomous /goal runs via AGENTS.md persona injection (gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna, gpt-5.5). Triggers on 'delegate to Codex', 'Codex subagent', 'code review agent', 'security audit', 'refactor with Codex', 'goal run', 'autonomous goal', 'have Codex weigh this'.