Fable Advisor
Escalate hard calls to a Fable subagent, with a live main-model probe
Test report
- Verdict
- Works with setup
- Score
- Tested
- Aug 5, 2026
- Environment
- Claude Code 2.x (agent harness)
⚠ This skill is no longer available upstream. Our re-check on Aug 10, 2026 couldn't find it any more (repo unreachable/deleted). The test below is what we measured on Aug 5, 2026 and we're leaving it up as a record — but there is nothing left to install, so we've removed the command.
The bundled current-model.sh probe is the part that can actually be measured, and it holds up: across a seven-case synthetic transcript matrix it answered 7/7 — mid-session /model switch, automatic Opus-to-Sonnet fallback, a poisoned tool result quoting a fake "Set model to haiku" line, a subagent sidechain leaking a different model, opusplan dual mode, and a <synthetic> placeholder tag — where the obvious one-liner an agent writes unaided (last assistant message's .message.model) scored 3/7 and returned literal garbage on two of them. Run live in a real session it correctly printed 'opus'. The friction is that SKILL.md hardcodes the probe at ~/claude-skills/fable-advisor/scripts/current-model.sh, a path from the author's own machine that does not exist after a normal ~/.claude/skills install, so step 0 fails until you add the symlink the install command below includes. The escalation doctrine itself is unmeasured here — exercising it means spending real Fable subagent tokens, and it needs Fable access on an Anthropic API plan, not Bedrock, Vertex or Foundry.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 6/10
- Docs & honesty 3/5
What Fable Advisor does
A delegation doctrine for Opus-main Claude Code sessions: Opus orchestrates, Sonnet subagents do the volume work, and a Fable subagent is consulted only at genuine decision points — before committing to a non-obvious approach, when stuck, or before declaring a task done. Opens with a bundled shell probe that derives the session's current main model from the transcript instead of trusting self-image, and routes to its sibling fable-coordinator skill on a Fable-main session. Requires Fable access on an Anthropic API plan.
How to install Fable Advisor
Nothing to install: the source repository no longer has this skill. If the author brings it back, our daily re-check will pick it up and the command will reappear here.
Commands — how to trigger Fable Advisor
-
/fable-advisorEscalate hard calls to a Fable subagent, with a live main-model probe
It also activates on plain-language prompts like these:
-
I'm about to commit to this architecture, is it the right call? -
Three fixes in and the test still fails, what should I do now? -
How should I split this migration work across subagents?
Frequently asked questions
- Is the Fable Advisor skill free?
- Yes. The skill itself is free from echo-of-machines/fable-advisor. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Fable Advisor work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Aug 5, 2026. Verdict: Works with setup. The bundled current-model.sh probe is the part that can actually be measured, and it holds up: across a seven-case synthetic transcript matrix it answered 7/7 — mid-session /model switch, automatic Opus-to-Sonnet fallback, a poisoned tool result quoting a fake "Set model to haiku" line, a subagent sidechain leaking a different model, opusplan dual mode, and a <synthetic> placeholder tag — where the obvious one-liner an agent writes unaided (last assistant message's .message.model) scored 3/7 and returned literal garbage on two of them. Run live in a real session it correctly printed 'opus'. The friction is that SKILL.md hardcodes the probe at ~/claude-skills/fable-advisor/scripts/current-model.sh, a path from the author's own machine that does not exist after a normal ~/.claude/skills install, so step 0 fails until you add the symlink the install command below includes. The escalation doctrine itself is unmeasured here — exercising it means spending real Fable subagent tokens, and it needs Fable access on an Anthropic API plan, not Bedrock, Vertex or Foundry.
- What is the Fable Advisor SkillProof Score?
- 7.6/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 6/10, docs & honesty 3/5.
- How do I install Fable Advisor?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Fable Advisor with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.