Menu
Iteratively fix test failures until all tests pass. Use when tests are failing and you want Claude to automatically plan and fix them in a loop.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Write in-game changelog updates as terminal documents in mansion.rec. Use when asked to write a changelog, add a terminal update, document recent changes, or catch users up on new features. Creates INBOX entries for the office terminal.
Add synonyms to verb word lists. Use when adding verbs/synonyms for MUD actions like look, touch, attack, use, take, open, close, or drop.
Build Discord bots using discord.py. Use when the user asks about Discord bot development, discord.py library, bot commands, slash commands, Discord intents, or Discord API integration. Covers Client, Bot, events, commands extension, app_commands, views, buttons, modals, and cogs.
Style conventions for writing entity DescriptionShort templates in mansion.rec. Use when adding or editing entities in data/worlds/mansion.rec to ensure consistent prose flow in /look output. Covers capitalization rules and punctuation for top-level entities vs container contents.
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.
Discord-specific markdown syntax for formatting message strings. Use when constructing strings that will be sent as Discord messages, including: text formatting (bold, italic, underline, strikethrough, spoilers), headers, subtext, code blocks, block quotes, lists, and masked links.
Based on SOC occupation classification
| name | healtests |
| description | Iteratively fix test failures until all tests pass. Use when tests are failing and you want Claude to automatically plan and fix them in a loop. |
| disable-model-invocation | true |
You are in a test-healing loop. Goal: make all tests pass.
just testq
If tests pass: Report success. The healing loop is complete.
If tests fail: Continue to Step 3.
EnterPlanMode toolCRITICAL: Your plan file MUST end with this exact section:
## Verification (Healing Loop)
This fix is part of a `/healtests` healing loop. After implementing:
1. Run `just testq`
2. If tests pass, the healing loop is complete
3. If tests fail, use `EnterPlanMode` to plan the next fix
The loop continues until all tests pass.
This ensures the loop context survives context clearing.