execution-guardrails — Claude Code skill: rules and installation
This skill sets checks for reporting problems, batching minor warnings, and editing text safely. Its rules apply even when fable-mode’s staged loop is off.
npx skills add mrtooher/fable-mode --skill execution-guardrails -a claude-codeWhat execution-guardrails does
Before reporting a problem, the agent must confirm it by searching, reviewing a diff, running the code, or checking the source. A lack of evidence alone does not establish a fault. The same standard applies when the agent describes a capability limit: it must say what it tried and where it failed.
Minor concerns are counted across a run. At the default threshold of three, the agent presents them together; the user can choose another threshold. A confirmed concern serious enough on its own does not have to wait. For substring edits, the skill calls for word-boundary matching and a follow-up search for malformed words. If a target appears more than once, the agent must narrow the match rather than replace every occurrence.
These rules apply to every task, with or without fable-mode. To install in Claude Code, run npx skills add mrtooher/fable-mode --skill execution-guardrails -a claude-code. The skill file is placed at ~/.claude/skills/execution-guardrails/SKILL.md.
What it helps with
- confirm a defect before reporting it
- batch minor concerns into one warning
- confine text replacements to word boundaries
- check edited files for malformed compound words
Example prompts for the agent
- Check that this issue is present before you report it.
- Keep a count of minor concerns and raise them together at three.
- Replace this text safely, then look for damaged neighboring words.
How to install execution-guardrails
1Skills CLI (Vercel)
npx skills add mrtooher/fable-mode --skill execution-guardrails -a claude-code2Manually: put the files in place
~/.claude/skills/execution-guardrails/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/execution-guardrails/
1Skills CLI (Vercel)
npx skills add mrtooher/fable-mode --skill execution-guardrails -a codex2Manually: put the files in place
.agents/skills/execution-guardrails/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add mrtooher/fable-mode --skill execution-guardrails -a cursor2Manually: put the files in place
.cursor/skills/execution-guardrails/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install mrtooher/fable-mode execution-guardrails2Manually: put the files in place
.github/skills/execution-guardrails/SKILL.md1Skills CLI (Vercel)
npx skills add mrtooher/fable-mode --skill execution-guardrails -a gemini-cli2Platform CLI
gemini skills install https://github.com/mrtooher/fable-mode --path execution-guardrailsManually: put the files in place
.devin/skills/execution-guardrails/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add mrtooher/fable-mode --skill execution-guardrails -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about execution-guardrails
How do I install execution-guardrails in Claude Code?
Run npx skills add mrtooher/fable-mode --skill execution-guardrails -a claude-code. The skill file is installed at ~/.claude/skills/execution-guardrails/SKILL.md.
What does execution-guardrails do?
It sets rules for confirming problems, grouping minor warnings, and making safer substring edits. The rules apply even when fable-mode’s staged process is not running.
When does the skill surface minor concerns?
At the default threshold of three, it reports accumulated minor concerns together. The user can set a different threshold, while a serious confirmed issue can be raised immediately.
