Skip to content
Get packs

Measure Before You Fix

Time a slow or stale step with a stopwatch before changing any timeouts or code.

Source Download skill
Repository
garrytan/gbrain
Stars
31,447
Licence
MIT
Files
1 file

What it does

When an alert says something is slow, stale, timed out, or stuck, this skill makes the agent measure the suspect step first instead of immediately raising a timeout or rewriting a pipeline. It guides the agent to run the step on the specific named item, compare the measured time against the alert's configured thresholds, and separate 'needs more time' from 'is wedged', since those need opposite fixes. It is designed for gbrain's health checks and sync monitors, and it changes no code itself.

Great for

  • Developers maintaining gbrain sync, doctor, or cron monitoring who get repeated timeout or freshness alerts
  • Engineers who want to avoid guessing at fixes for slow background jobs
  • Teams that want a documented measure-first triage routine for temporal alerts

Skip it if

  • Projects not using gbrain, since the thresholds and commands are specific to its tooling and would not apply
  • Urgent production outages where the agent should act on known runbooks rather than spend time measuring; the skill is meant to delay structural fixes, not accelerate them
  • Anyone expecting it to fix bugs automatically, because the skill is explicitly read-only and only produces a measurement verdict

Try asking

  • “My gbrain sync keeps timing out, can you measure where the time goes before we raise the timeout?”
  • “Doctor is showing a sync freshness warning, check whether it is real against the 24h and 72h thresholds”
  • “The sync job seems wedged, figure out whether it needs more time or is actually stuck”
  • “Our cron monitor keeps paging about stale data, check whether its threshold is below the doctor's warn line”
Step-by-step guide: how to use Measure Before You FixInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    measure-before-you-fix.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip measure-before-you-fix.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip measure-before-you-fix.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the agent the exact alert text and the named source so it can measure the specific entity rather than the whole brain
  • Let the agent search the brain for prior incidents of the same alert first, since recurring alerts often have a recorded verdict
  • Review any proposed fix after the measurement; the skill intentionally leaves the actual change to a separate step

What's inside

  • SKILL.md

Redistributed unmodified from garrytan/gbrain under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.