agentleFS
syangasyanga/astack/skills/claude-restore-skills/SKILL.md

claude-restore-skills

Skill0 starsChanged 57 days agoMITView on GitHub

Claude Code hook that restores the text of user-invoked skills after a compaction or resume drops it from context.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    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:

    After a compaction, Claude Code sometimes attaches its own copy of a skill, cut at 20,000 characters.

    line 27

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Restore user-invoked skills after compaction or resume

scripts/hook.py is a Claude Code SessionStart hook for the resume and compact sources. It reads the session transcript and finds each skill the user invoked with a slash command. For each one whose text is no longer in context, it adds the skill's current SKILL.md text and the invocation's arguments. When the file is gone, it uses the text recorded at invocation. It skips skills the model invoked through the Skill tool, because the model can invoke those again.

The hook also skips invocations on branches abandoned by a rewind. When the first entry after a rewind has a parent missing from the transcript, the hook cannot tell the abandoned branch from lost entries. It can then restore that branch's skills.

The hook splits its output into parts. The comment on PARTS in scripts/hook.py says why. When the skills do not fit in all parts together, the hook leaves out the oldest skills until the rest fit and lists their files and arguments. If the newest skill alone does not fit, the hook cuts it and points to its file.

After a compaction, Claude Code sometimes attaches its own copy of a skill, cut at 20,000 characters. It attaches that copy after the hook runs, so the skill arrives twice. On a later resume, the hook skips a skill whose copy in context is complete.

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.