agentleFS
sabirmgdsabirmgd/terminal-first-goal-loop/skills/hot-reload/SKILL.md

hot-reload

Skill2 starsChanged 2 months agoMITView on GitHub

Drive debugging and feature completion through the smallest failing boundary first, reuse valid evidence, and keep the full journey as the last step. Use when repeated rebuilds, redeploys, or end-to-end reruns are slowing the task down.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Keep the full browser, end-to-end, or release journey for last.

    line 22
  2. Distinctive:

    If the same defect class comes back twice, stop local patching and escalate to a deeper design or invariant review.

    line 23
  3. Distinctive:

    When a change touches only one boundary, do not throw away unrelated green proof.

    line 40
  4. Distinctive:

    Only invalidate evidence that the change can actually affect.

    line 41
  5. Distinctive:

    Stop this loop and switch to a wider review when:

    line 52

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Hot Reload

This skill changes the order of proof, not the definition of done.

The Loop

Repeat this loop until the boundary is green:

exact failure -> smallest failing check -> smallest coherent fix -> same check again -> next boundary

Rules

  1. Freeze the exact failing boundary before editing.
  2. Pick the cheapest check that can prove the current claim.
  3. Reuse accepted evidence unless the change actually invalidates it.
  4. Move up one boundary at a time.
  5. Keep the full browser, end-to-end, or release journey for last.
  6. If the same defect class comes back twice, stop local patching and escalate to a deeper design or invariant review.

Boundary Ladder

Typical order:

  • static check or type check
  • unit test
  • integration or contract test
  • hot-reload request
  • slot or sandbox proof
  • full browser or consumer journey
  • release or deployed-runtime proof

Evidence Reuse

When a change touches only one boundary, do not throw away unrelated green proof. Only invalidate evidence that the change can actually affect.

A useful record after each step is:

  • what failed
  • what changed
  • what proof stayed valid
  • what proof must be rerun next

Stop Conditions

Stop this loop and switch to a wider review when:

  • the same root defect recurs twice
  • the next fix would cross a larger architectural boundary
  • the environment is the real blocker rather than the code
  • the task now needs the final browser or release proof

Output Contract

HOT RELOAD | boundary=<summary> | level=<name>
FAILURE | <exact failing signal>
DELTA | <what changed>
REUSED | <still-valid proof>
NEXT | <smallest next check>
FULL JOURNEY | allowed|not-yet

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.