Index of every page on this site: llms.txt
All skills
expo avatar

/llp-orient

@f428a37
by expoexpo/exact4.7k stars
132
Run it yourself

Before editing a subsystem, file, or starting a task, assemble the governing LLP context — read the root LLP, follow the @ref annotations on the code in scope, and find the LLPs covering the relevant systems. Produces a compact context pack of the constraints a change must respect. Read-only.

Requires /llp-adopt
  • 1 file
  • 3.6 KB
  • Updated 2 weeks ago
  • GitHub
Use this Skill: https://skilld.dev/gh/expo/exact/llp-orient

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈77 tokens for metadata: the name and description. ≈825 when used: this file.

Description uses 3.8% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈306 characters, or 3.8%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

llp-orient

<!-- @ref https://github.com/ccheever/llp/blob/v0.5.2/llp/0008-distributed-agent-skills.rfc.md#the-skills--core-five-plus-flags — skill contract --> <!-- @ref https://github.com/ccheever/llp/blob/v0.5.2/llp/0000-linked-literate-programming.explainer.md#2-reference-syntax — the grammar this reads -->

Load the design rationale governing the area you're about to touch, before touching it. This is the consumption side of LLP — orienting first is what turns @ref from a possibility into a habit.

Invoke as /llp-orient <path-or-task> (file, directory, subsystem, or task description). With no argument, use the current change scope (git diff --name-only).

Trigger

Start of a coding task that touches an area with documented design. The orient-first rule lives in the repo's always-loaded AGENTS.md; this skill is the detailed form that rule routes to for non-trivial work — don't rely on ambient description-match alone.

Invariants

  • MUST be read-only: never edit code or LLP documents.
  • MUST scope to the task: pull the constraints that bind this change, not a corpus summary.
  • MUST cite, not restate: reference policy as LLP NNNN#anchor rather than paraphrasing sections.
  • MUST NOT invent constraints. If something isn't documented, say so.
  • MUST NOT guess the rationale behind a broken @ref — report it and hand off to llp-maintain.

Workflow

Recipe (advisory) — read llp/foundation/ first (the kernel; the root LLP is there once Active — on a fresh adoption it is still Draft and sits in current/), then skim llp/current/ (the declared working set); both hold relative links — symlinks, or small files containing the relative path — so ls answers and either form resolves. Then the root LLP's systems vocabulary. Grep '@ref' over the files in scope and Read each target — these are the highest-signal context. Grep the **Systems:** headers across llp/ for LLPs governing the subsystems touched; read the governing ones (Spec, Decision, Principles, relevant RFCs). Follow Related: and nearby @ref targets one hop out; stop there unless something is clearly load-bearing. Stale-pin nudge (skip offline): read the installed tag from the skills receipt (or a source: pin), compare git ls-remote --tags <origin> — if upstream is newer, say so in one line ("skills at v0.3.0, upstream v0.5.2 — run /llp-adopt update") and move on; never update from here.

Artifact

A context pack with these sections, faster to read than the documents it points at:

  • Scope — files / subsystems queried
  • Governing LLPs — numbered list, one-line gloss each; mark which are in foundation/ and which are in current/ (in-flight — its guidance may still move)
  • Active constraints — bullets, each citing LLP NNNN#anchor
  • Relevant rationale — the why, same citation form
  • Open questions / risks for this area
  • Likely doc updates if the change lands
  • Hand-off — one of none | llp-adopt | llp-create | llp-review | llp-maintain

Hand-offs

  • No corpus (llp/ missing or stub) → llp-adopt.
  • The task needs a decision no LLP covers → llp-create (or llp-review if proposal-shaped).
  • An @ref target or anchor doesn't resolve → llp-maintain.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No third-party reports yet.

Signed by skilld at f428a37. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 weeks ago
source
ccheever/llp@v0.5.2

README badge

README badge for expo/exact/llp-orient