agentleFS
ziya-aiziya-ai/ziya/.agents/skills/hot-patch-static-assets/SKILL.md

hot-patch-static-assets

Skill42 starsChanged 3 months agoMITView on GitHub

Test changes to frontend/static assets served by a long-running server you cannot restart, by locating the actual on-disk serve path and overwriting bundles in place.

  • Installs packages

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Identify the actual serving path: Don't assume the server reads from your repo.

    line 25
  2. Very specific:

    Check for installed-package paths (site-packages/, node_modules/, dist/)

    line 28
  3. Very specific:

    Look for non-editable pip install layouts — source edits won't propagate

    line 29
  4. Distinctive:

    Read the server's static-mount / template-loader code

    line 31
  5. Distinctive:

    Trigger a fresh request: No restart needed for static files — the next request reads the new bytes off disk.

    line 40

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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-Patching a Running Server's Static Assets

When to use

You edited frontend/static code (JS bundle, HTML template, CSS, image), rebuilt it, and your changes don't appear at runtime — yet you can't restart the server (or restarting is expensive/disruptive).

Core insight

A running server's process memory (Python/Node code) is frozen until restart, but static files it serves from disk can be swapped live. Each new request re-reads them.

Procedure

  1. Identify the actual serving path. Don't assume the server reads from your repo. Find where the running process loads assets from:
    • lsof -p <pid> to see open files
    • Check for installed-package paths (site-packages/, node_modules/, dist/)
    • Look for non-editable pip install layouts — source edits won't propagate
    • Grep server logs/config for template or static directories
    • Read the server's static-mount / template-loader code
  2. Build the asset locally (npm run build, craco build, etc.) into the project's normal output directory.
  3. Copy build output to the actual serve path — even if it's outside your usual writable scope (e.g. ~/.pyenv/.../site-packages/...). Try the write; many sandboxes allow it.
  4. Trigger a fresh request. No restart needed for static files — the next request reads the new bytes off disk.
  5. Verify with the rendering/test tool, not by inspecting the build output. Caches (browser, service worker, CDN) can mask success; if results look stale, hard-reload or bust the cache.

Red flags that you're hitting this problem

  • "I rebuilt but the behavior is identical"
  • The error message is byte-identical to before your fix
  • grep confirms your code is in the new bundle, but runtime doesn't reflect it
  • Project was pip install'd (not pip install -e) or assets are vendored into a package
  • Two copies of templates/ or static/ exist on disk

Doesn't work when

  • The change is in server code (Python/Node) — needs process restart
  • Assets are bundled into memory at server startup (read-once)
  • A reverse proxy/CDN caches aggressively without a bust mechanism
  • The server uses a hashed manifest and your new bundle has a different hash that nothing references

Mnemonic

"Find where it's actually served from, drop the new bundle there, re-request."

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.