llp-adopt

星标13

分支1

更新时间2026年7月23日 18:00

Bring LLP to a repository and keep its installed skills current — the single entry point for "set up LLP here" and "update the LLP skills". Scaffold mode (fresh structure), retrofit mode (survey the codebase, draft a provenance-tagged root LLP, propose an adoption plan), install step (copy skills from a tagged release with a receipt), and update mode (verified, diff-approved skill updates).

安装

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

来源

ccheever

ccheever/llp

打开 GitHub 仓库查看创作者相关仓库

下载

llp-adopt

One entry point for "add LLP to this project" and "update the LLP skills". Invoke as /llp-adopt (auto-detect scaffold/retrofit), --mode scaffold / --mode retrofit, or /llp-adopt update.

Trigger

"Set up / bring LLP to this repo," for any repository state; "update the LLP skills," in any repo with installed skills.

Invariants — adoption

MUST extend an existing llp/ tree if one exists — never start a second setup layer.
MUST write only new files and clearly-marked managed blocks; existing user prose is never overwritten without showing a diff and asking. Reruns MUST be idempotent (no duplicate files or blocks).
MUST keep every generated document Draft and provenance-tagged: claims are [observed] (say where), [confirmed] (named human + date), or [inferred] — and [inferred] claims never survive into a document proposed for promotion.
MUST propose @ref annotations and apply them only after approval — no mass annotation, and none on volatile early code.
MUST NOT create a batch of speculative LLPs in either mode; establish orientation and the loop, then hand off.
When a tool expects CLAUDE.md, symlink it to AGENTS.md when safe rather than maintaining divergent copies.

Invariants — install and update (LLP 0010 §1)

MUST install only from a tagged release (the tag's tree is the release), as copies never symlinks, and write a receipt beside the installation: origin, tag, resolved commit, profile flags (core; review-plus; delivery, which also records the repo's delivery policy in repo-local config), and a per-file content hash. The receipt — not source: frontmatter — is the trust root for updates.
MUST confirm before writing user/global skill directories (they affect every repo for that user); project-local may proceed more freely.
Update MUST three-way-check each file against the receipt hash. Any mismatch is a local fork: abort the coupled update, never clobber, and offer discard (restore managed copy) or detach (file leaves the managed set permanently; receipt records it; later updates ignore it and note upstream changes without applying).
Update MUST present the full changeset for the installed profile as defined at the target tag as one diff, revalidate the approved state immediately before mutation, and be recoverable to either endpoint (the receipt write is the commit point). Runs are single-flight per installation.
If upstream is unreachable, a tag moved or vanished, or content can't be verified against the resolved commit: report exactly that and write nothing. Cannot-verify never degrades to guess.

Workflow

Recipe (advisory) — Adoption: inspect first (existing llp/? agent files? how much code?): little to mine → scaffold; substantial history → retrofit. Both share the spine: llp/, root LLP 0000 (Explainer, **Role:** Root), the AGENTS.md managed block. Retrofit adds: survey, draft root LLP, propose 2–5 subsystem LLPs, layered annotation plan (LLP 0001). Install: offer after the managed blocks; default profile core, flags opt-in. Update: fetch the tag's tree (shallow clone or archive; git required), diff, approve, apply as a staged set with prior bytes retained until the receipt lands; rollback is git revert where the installation is committed, the backups otherwise. Releasing (this repo): bump source: pins and pinned URLs → commit → tag → push tag — a skill-file change is not installable until tagged.

Artifact

Scaffold: llp/, an LLP 0000 skeleton, the managed blocks. Retrofit: those plus a survey summary, Draft provenance-tagged document(s), and a proposed adoption plan. Install/update: pinned skill copies, the receipt, and (with the delivery flag) the repo policy file.

Hand-offs

Corpus exists → daily work starts with llp-orient; author follow-on docs with llp-create.
Ratify generated drafts with llp-review where stakes warrant; keep healthy with llp-maintain — whose audit notes stale pins and hands back to /llp-adopt update.

同仓库更多 Skills

同仓库

llp-super-refine

ccheever/llp

Author-invoked dual-family refine loop for one or more Draft/Review LLPs — each round, two fresh mutually-blind reviews from different model families, orchestrator revision, verdict-lined artifacts — until both families mark the same revision READY, a delta round finds no MATERIAL concern in pre-existing text across families, or a bounded stop rule escalates to the author. Use when the user invokes /llp-super-refine or asks to refine an LLP until independent models agree it is ready.

2026-07-2413

llp-create

ccheever/llp

Create a new LLP document with the next available number, the NNNN-slug.type.md filename convention, and a scaffolded metadata header — or extend an existing LLP when one already covers the topic.

2026-07-2313

llp-maintain

ccheever/llp

Keep an LLP corpus in sync with its code. Detects code/doc drift, validates @ref annotations interactively, checks provenance tags, and proposes reconciliations, ref repairs, and status changes. Modes — pre-pr, audit, reconcile, retire-proposal. Proposes; never applies.

2026-07-2313

llp-orient

ccheever/llp

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.

2026-07-2313

llp-review

ccheever/llp

Run the LLP 0005 review loop for an LLP document — orchestrate reviews from other model families, write one provenance-tracked artifact per review actually received under llp/reviews/, and never fabricate a review or accept on the author's behalf. Review intensity scales with stakes; the author decides when it's enough.

2026-07-2313

llp-ship

ccheever/llp

Take one or more tasks (Linear tickets, ticket files, accepted LLPs, free text) end to end — plan lanes, implement in run-namespaced worktrees, integrate on a run branch, verify and review the exact integrated commit, deliver per repo policy (direct push by default), and clean up without destroying recoverable work. Use when the user invokes /llp-ship or hands over tasks to be implemented and shipped.

2026-07-2313

name	llp-adopt
description	Bring LLP to a repository and keep its installed skills current — the single entry point for "set up LLP here" and "update the LLP skills". Scaffold mode (fresh structure), retrofit mode (survey the codebase, draft a provenance-tagged root LLP, propose an adoption plan), install step (copy skills from a tagged release with a receipt), and update mode (verified, diff-approved skill updates).
source	ccheever/llp@v0.3.0

llp-adopt

One entry point for "add LLP to this project" and "update the LLP skills". Invoke as /llp-adopt (auto-detect scaffold/retrofit), --mode scaffold / --mode retrofit, or /llp-adopt update.

Trigger

"Set up / bring LLP to this repo," for any repository state; "update the LLP skills," in any repo with installed skills.

Invariants — adoption

MUST extend an existing llp/ tree if one exists — never start a second setup layer.
MUST write only new files and clearly-marked managed blocks; existing user prose is never overwritten without showing a diff and asking. Reruns MUST be idempotent (no duplicate files or blocks).
MUST keep every generated document Draft and provenance-tagged: claims are [observed] (say where), [confirmed] (named human + date), or [inferred] — and [inferred] claims never survive into a document proposed for promotion.
MUST propose @ref annotations and apply them only after approval — no mass annotation, and none on volatile early code.
MUST NOT create a batch of speculative LLPs in either mode; establish orientation and the loop, then hand off.
When a tool expects CLAUDE.md, symlink it to AGENTS.md when safe rather than maintaining divergent copies.

Invariants — install and update (LLP 0010 §1)

MUST install only from a tagged release (the tag's tree is the release), as copies never symlinks, and write a receipt beside the installation: origin, tag, resolved commit, profile flags (core; review-plus; delivery, which also records the repo's delivery policy in repo-local config), and a per-file content hash. The receipt — not source: frontmatter — is the trust root for updates.
MUST confirm before writing user/global skill directories (they affect every repo for that user); project-local may proceed more freely.
Update MUST three-way-check each file against the receipt hash. Any mismatch is a local fork: abort the coupled update, never clobber, and offer discard (restore managed copy) or detach (file leaves the managed set permanently; receipt records it; later updates ignore it and note upstream changes without applying).
Update MUST present the full changeset for the installed profile as defined at the target tag as one diff, revalidate the approved state immediately before mutation, and be recoverable to either endpoint (the receipt write is the commit point). Runs are single-flight per installation.
If upstream is unreachable, a tag moved or vanished, or content can't be verified against the resolved commit: report exactly that and write nothing. Cannot-verify never degrades to guess.

Workflow

Recipe (advisory) — Adoption: inspect first (existing llp/? agent files? how much code?): little to mine → scaffold; substantial history → retrofit. Both share the spine: llp/, root LLP 0000 (Explainer, **Role:** Root), the AGENTS.md managed block. Retrofit adds: survey, draft root LLP, propose 2–5 subsystem LLPs, layered annotation plan (LLP 0001). Install: offer after the managed blocks; default profile core, flags opt-in. Update: fetch the tag's tree (shallow clone or archive; git required), diff, approve, apply as a staged set with prior bytes retained until the receipt lands; rollback is git revert where the installation is committed, the backups otherwise. Releasing (this repo): bump source: pins and pinned URLs → commit → tag → push tag — a skill-file change is not installable until tagged.

Artifact

Hand-offs

Corpus exists → daily work starts with llp-orient; author follow-on docs with llp-create.
Ratify generated drafts with llp-review where stakes warrant; keep healthy with llp-maintain — whose audit notes stale pins and hands back to /llp-adopt update.