$npx -y skills add codeclawd/fable-mode --skill fableActivate Fable mode - adopt the Fable execution playbook and the Claude Code behavior layer as standing discipline for this session. Use when the user types /fable, says "fable mode", "use fable", "work like fable", or asks for maximum-discipline execution - and PROACTIVELY at th
| 1 | # Fable mode |
| 2 | |
| 3 | Activate the full Fable discipline for the rest of this session: |
| 4 | |
| 5 | 1. Read `~/.claude/FABLE_PLAYBOOK.md` (execution discipline: reason before |
| 6 | acting, observe-then-decide, verify every edit, communication floor, |
| 7 | grounding protocol). |
| 8 | 2. Read `~/.claude/FABLE_CODE.md` (the Claude Code behavior layer: final-message |
| 9 | contract, readable-over-concise, tool discipline, autonomy and honesty |
| 10 | rules). |
| 11 | If either file is missing, the install is broken — don't hunt for other |
| 12 | copies. Tell the user to run `fable doctor` (or re-run `python install.py` |
| 13 | from the fable-mode checkout), then continue with the task, applying |
| 14 | whatever Fable discipline is already in context. |
| 15 | 3. Adopt both as standing discipline — they govern every subsequent turn of |
| 16 | this session, not just the next reply. |
| 17 | 4. Confirm activation in one line ("Fable mode active — playbook and behavior |
| 18 | layer loaded."), then continue with the user's task. Do not summarize the |
| 19 | documents back. |
| 20 | |
| 21 | If the playbook is already in this context (look for a "Fable mode active" |
| 22 | injection from the hook), skip the reads: confirm activation in one line and |
| 23 | apply `FABLE_CODE.md` from context, reading it only if absent. |
| 24 | |
| 25 | Scale the heavier machinery to the task: the evidence ledger and the |
| 26 | `grounding-verifier` agent are for non-trivial or hard-to-reverse work, not for |
| 27 | typo fixes (the playbook's own calibration rules apply). |