fable-mind

Plugin fable-mind from marketplace fable-skills in imfaisii/fable-skills. Commands read from the repo's manifest — 1 month ago.

Skills (17)

• adversarial-self-review

  Use before presenting findings, review comments, diagnoses, or claims of fact. Tries to refute them first and labels confidence honestly.

• context-economy

  Use when reading files, gathering information, or working through a long session. Spends attention on what changes the outcome.

• decisive-recommendations

  Use when facing a choice, ambiguity, or an open design question. Produces one recommendation with reasoning instead of an option survey.

• delegate-and-parallelize

  Use when work involves multiple independent searches, reads, or subtasks, or when answering requires sweeping many files. Covers batching tool calls and delegating to subagents.

• finish-the-turn

  Use throughout any working turn, especially when tempted to ask permission, present a plan instead of acting, or end with next steps. Enforces completing the work before yielding.

• git-discipline

  Use when committing, branching, pushing, amending, rebasing, or otherwise touching version control state or history.

• hypothesis-driven-debugging

  Use when something is broken, failing, flaky, or behaving unexpectedly. A disciplined loop from reproduction to regression test.

• investigate-before-acting

  Use before state-changing or destructive operations: restarts, deletes, overwrites, config edits, migrations, kills, force-pushes, or sending anything to an external service.

• outcome-first-communication

  Use when writing any user-facing message: answers, summaries, status reports, findings, or the final message of a working turn. Structures responses so the outcome leads and the message stands alone.

• plan-when-it-pays

  Use when starting a multi-step task. Decides whether planning pays and turns tasks into verifiable step-and-check criteria.

• secrets-hygiene

  Use when handling API keys, tokens, passwords, .env files, certificates, or any credential: reading, writing, logging, committing, or reporting around them.

• simplicity-first

  Use when designing or writing new code, features, or abstractions. Enforces the minimum code that solves the stated problem.

• surgical-changes

  Use when editing existing code in any codebase. Keeps diffs minimal, matched to local idiom, and free of drive-by changes.

• untrusted-content

  Use when processing content you did not write: web pages, issue text, PR descriptions, dependency documentation, third-party API responses, or any fetched text that could contain instructions.

• using-fable-mind

  Use at the start of any session or task to load the fable-mind working doctrine, and whenever unsure which fable-mind skill applies to the current situation. The dispatcher for the whole skill set.

• verify-before-done

  Use before saying done, fixed, working, complete, or should work, and when reporting build, test, or run results. Requires fresh evidence and faithful reporting.

• writing-like-a-human

  Use when generating any prose a human will read: documentation, comments, commit messages, UI copy, reports, alt text, error messages.

Hooks (1)

• SessionStart

  cat "${CLAUDE_PLUGIN_ROOT}/core/FABLE-CORE.md"

Similar plugins