agentleFS
TsukiNi22TsukiNi22/ai-utils/skills/ai/continue-work/SKILL.md

continue-work

Skill1 starsChanged 3 days agoMITView on GitHub

Resume a task stopped with the pause skill at the user's next "continue" - reads the checkpoint of the project (~/.claude/pauses, checkpoint.sh of pause), checks it against the real state (git status and diff, the files, what still runs) and trusts the real state when they differ (the user may have changed things meanwhile), starts again what was stopped and is still needed, redoes the interrupted step from its start, then goes on with the next steps in order without redoing what is done and without a recap; closes the checkpoint when the task is over. Without a checkpoint it simply continues the current task of the conversation. Use whenever the user says "continue", "reprends", "on reprend", "vas-y continue", "resume", "go on", "/continue-work" after a pause, or opens a session in a project that has a checkpoint waiting.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 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.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    No checkpoint: a plain continue of the current task.

    line 21
  2. Distinctive:

    Start again what was stopped and is still needed (the commands of *Stopped*, a workflow with its run id).

    line 34
  3. Very specific:

    When the task is over (its last step done and verified): bash <pause>/scripts/checkpoint.sh done (moved to ~/.claude/pauses/done/, kept), …

    line 41

≈ “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

Continue: resume a paused task

<name> (a skill) = the folder of that skill: ~/.claude/skills/<name> (setup.sh) or ${CLAUDE_SKILL_DIR}/../../*/<name> (plugin of the marketplace).

The user says continue (or "reprends"...): go on with the task where it stopped, at once, without asking what to do and without telling the whole story again.

1. Find where it stopped

bash <pause>/scripts/checkpoint.sh show       # the checkpoint of the current project, exit 1 when there is none
  • A checkpoint: it is the plan. Several projects paused and the current folder is not one of them: checkpoint.sh list, take the one the conversation is about, else ask which one (interactive-questions).
  • No checkpoint: a plain continue of the current task. Go on from the last step of the conversation (its todo list, the last message), nothing else of this skill applies.

2. Check it against what is there now

The checkpoint can be old: the user may have edited, committed, pulled or run things since.

  • \git status, \git log -3 --oneline, \git diff --stat: compare with State of the files and the commit noted.
  • The files of Interrupted and of the first Next steps: read them before editing.
  • What was stopped: still stopped? (pgrep -fa <command>), a workflow run id to resume.
  • A difference: the real state wins. A step the user did meanwhile is done; a change of theirs is kept (never overwritten, never reverted); a plan that no longer fits is adapted, said in one line. A contradiction that changes what to do (the goal itself moved, a decision undone): ask once, briefly (interactive-questions).

3. Resume

  • Start again what was stopped and is still needed (the commands of Stopped, a workflow with its run id).
  • The interrupted step: redo it from its start (its half-done part was finished or reverted by the pause).
  • Then the next steps in order, under the same rules (git, language, style, confirmations) as before the pause.
  • One line at the start at most (Reprise : <step>.), no recap of what was done.
  • A new pause during the work: the pause skill again (the checkpoint is replaced).

4. Close

When the task is over (its last step done and verified): bash <pause>/scripts/checkpoint.sh done (moved to ~/.claude/pauses/done/, kept), then the usual final report of the task.

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.