agentleFS
kaotyprkaotypr/context-circuit/.agents/skills/cc-complete/SKILL.md

cc-complete

Skill0 starsChanged 17 days ago0BSDView on GitHub

Mark Context Circuit plans completed and reconcile the durable product knowledge those plans changed, which together finish the request.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write the note in the words the person used, including their language.

    line 17
  2. Distinctive:

    Keep implementation-specific evidence in the plan, and keep those plans' own identifiers out of every note reconciled from them.

    line 45
  3. Very specific:

    When the result carries archive_offer, the person has that many finished records of their own not archived.

    line 56

≈ “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

Mark a plan completed

Delivery does not mark a plan done. Completion happens when a person explicitly asks for it, and it is two acts, not one: append a short completion note to each record, and reconcile the durable product knowledge those plans changed. The request is finished only when both are done.

sh <root>/.context-circuit/cli --json record complete --id p0001 --text 'User requested completion.'

Write the note in the words the person used, including their language. It records that someone asked for this, so a translation of it is a paraphrase of evidence rather than a tidier form of it. The knowledge reconciled afterwards is English regardless.

The command stamps completed_at in the plan's frontmatter. Its result carries completed, the ID of the plan it marked, the catalog entries scoped to that plan's repositories, and reconcile_required naming the judging whenever there are any. In a Git workspace it also carries commit_recommended, the plan, intent, and ID files left uncommitted: ask the person once whether to commit and push them, or commit only. Their answer authorizes that and nothing more; a refused push is handled as AGENTS.md says under delivery. Completion reporting success is the record written, not the request finished.

Judge, do not rewrite

The returned entries are a candidate set to judge, never a list to rewrite.

Most completions change no durable knowledge. Recording that in the completion note is the normal outcome, not a skipped step — and it is an assertion about the code, so make it after reading the entries rather than before.

Where meaning did change, edit the note and its catalog entry in one pass and move its reviewed date. A renamed or retired identifier is a glossary row. See .agents/skills/cc-knowledge/SKILL.md for what a note may say and how it anchors.

Completing several plans at once reconciles once across the set, not once per plan.

Keep implementation-specific evidence in the plan, and keep those plans' own identifiers out of every note reconciled from them. Run the diagnostic afterwards:

sh <root>/.context-circuit/cli check

Other work need not wait for unrelated knowledge updates.

Offer to archive

When the result carries archive_offer, the person has that many finished records of their own not archived. Once the request is finished, offer to archive them, and run record archive only on a yes; a no changes nothing. It moves only their own finished records and returns commit_recommended, which is asked about like any other: commit and push, or commit only, on their authorization.

sh <root>/.context-circuit/cli --json record archive

What completion is not

It does not remove a worktree, delete a branch, or deliver anything. Worktrees and branches remain after completion, and each of those is a separate act needing its own request.

Completing a plan does not mean its work was merged. When another approach replaces a plan, complete it anyway and say so in the note: the record should describe what happened, and the alternative is a plan left open forever.

One consequence needs care. record order treats any plan carrying completed_at as a satisfied dependency, so a plan that merged nothing still satisfies whatever depends on it. Move those dependencies to the plan that did the work.

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.