agentleFS
openvigletopenviglet/shio/.claude/skills/shio-content-files/SKILL.md

shio-content-files

Skill3 starsChanged 8 days agoApache-2.0View on GitHub

Read before touching Shio's content-as-files projection, the shio CLI, or blueprints. Triggers on shio/content/**, shio/folders/**, shio/assets/**, shio/post-types/*.json, content-files.mjs, content-sync.mjs, content-store.mjs, content-assets.mjs, content-notes.mjs, the three-way merge and content-base.json, sidecar files, movedFrom, shio pull/push/apply/verify/report/remember, blueprint.json, ShBlueprintRegistry, and the CI gate order.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    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. Very specific:

    a declared loss the format no longer has fails the build: If you add a normalization in content-files.mjs, content-notes.mjs or …

    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

Content as files, the CLI, and blueprints

Law P4 in practice: content projects to one Markdown file per post, and the agent's cheapest tools (Edit/Write/Grep) reach it.

Read the area file before editing:

  • docs/agents/content-files.md — the format, the three-way merge, sidecars, folders-as-files, moves, assets
  • docs/agents/cli.md — every shio command and the CI gate order
  • docs/agents/blueprints.md — the package format and its load-time validation

The rules that are load-bearing

  1. The path is the URL, not the folder chain. A file whose path and url: disagree is an error — a half-done rename is a move on one side and a create on the other.
  2. Only the DRAFT projects. publish: in a file is an instruction push forwards, never state pull writes back.
  3. moves is the document's one intent channel. State cannot express "this is the same post": without it, --prune deletes the old URL and creates a new post with a new id and no history. Ambiguity is refused; a plausible-but-unprovable pair is printed with the movedFrom: line to add.
  4. The merge happens in the projection's domain (SH253). Pass a remote document through projectedPost on the way in, or a normalization the format applies makes every pull report a local change nobody made — forever.
  5. A blueprint may not name its own site and may not prune.
  6. Losslessness beats prettiness, and a value a format cannot hold must be refused, not trimmed (SH256, SH257).

Guards that already exist

cli/test/content-round-trip.test.mjs asserts the round trip as a property over an adversarial value list × six round trips, with projectedPost a fixed point — and a declared loss the format no longer has fails the build. If you add a normalization in content-files.mjs, content-notes.mjs or content-assets.mjs, that is the test that will tell you what it costs. test/emit.test.mjs fails on blueprint/front-end drift and asserts the section overlay changes exactly three files.

Two things the CLI suite earned the hard way (SH198)

  • A fake filesystem normalizes its keys (test/support/paths.mjs) — the commands join with node:path, so on Windows a POSIX literal never matches and the test fails on a message several assertions from the cause. Do not "fix" it by making a command emit forward slashes.
  • A test that arms a live handle releases it in t.after, never only on the happy path.

Build order note: the CLI suite needs pnpm -C shio-client run build first; shio-sections' suite needs pnpm -C shio-model run build.

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.