agentleFS
stevehansenstevehansen/eidet/.claude/skills/portal/SKILL.md

portal

Skill4 starsChanged 54 days agoMITView on GitHub

Prime on Eidet's Portal before changing it — the per-repo generated HTML state view behind GET /api/eidet/portal, PortalRenderer's one-fetch section loop, the IPortalSection contract and its five v1 sections (Identity, Architecture, Procedures, Heuristics, Health), PortalMarkup citation anchors and escaping, and the augment=off liveness model. Use when the task touches a Portal section, portal rendering or citations, or the Web UI's portal view. Not for the approve-a-page curation loop (see canon), not for ranked agent-facing retrieval (see recall).

10 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

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

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 0

    Distinctive

    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:

    Only one test file covers this domain (PortalRendererTests).

    line 42

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “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

Portal — priming

Canonical spec: docs/domains/portal.md — read it for the section rules, all invariants, key files, and gotchas. Terms of record: UBIQUITOUS_LANGUAGE.md § Canon (for the Canon/Portal distinction).

A read-only projection of live memories into cited HTML sections: "what has this agent actually learned here?", auditable by a human. Human surface only — no MCP tool, on purpose.

Core invariants (get these right)

  • Every claim is a citation — render through PortalMarkup.Cite (anchor to #memory/<id> with data-mid). A section must never state something it can't hyperlink.
  • Escape everything via PortalMarkup.Esc; memory content is agent-written and unescaped.
  • Selection rules are deterministic and total — first-match precedence, ties broken on id, so two renders of the same corpus agree.
  • Sections read only PortalContext — the renderer pre-fetches once; a section issuing its own query breaks that contract.
  • v1 renders live and caches nothing; augment is hardcoded "off" (the field exists so the API shape survives augmentation landing).
  • Normalize the repo id once at the top — BrowseAsync normalizes internally, GetCountsByTypeAsync doesn't.

Key files / reuse

  • src/Eidet.Core/Portal/PortalRenderer.cs — the facade + the default section list.
  • src/Eidet.Core/Portal/IPortalSection.cs — implement this (it's internal) to add a section.
  • src/Eidet.Core/Portal/PortalMarkup.cs — escaping, citations, ordering; never hand-roll HTML.
  • src/Eidet.Core/Portal/PortalContext.cs — the only input a section gets.

Gotchas

  • The pre-fetch is bounded — a large repo renders a partial Portal with no warning.
  • AlwaysPresent is a promise the section keeps itself; the renderer only drops nulls.
  • Health metrics stick to what every memory carries (counts + a CreatedAt histogram).
  • The Glossary section doesn't exist yet and is where Canon convergence is designed to land — don't bolt canon:* reads onto an existing section.
  • Only one test file covers this domain (PortalRendererTests); a changed section rule needs a test in the same PR or nothing catches it.

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.