跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
开发与测试

ln-91-engineering-baseline-builder

Creates project technical rules and recurring local procedures from the actual stack; edits guidance only, not architecture targets or application code.

levnikolaevich · 已有静态检查记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add levnikolaevich/claude-code-skills --skill ln-91-engineering-baseline-builder

技能说明


name: ln-91-engineering-baseline-builder description: "Creates project technical rules and recurring local procedures from the actual stack; edits guidance only, not architecture targets or application code."

Engineering Baseline Builder

Goal: Create or reconcile approved technical guidance and local procedures. Edit only authorized documentation, agent instructions and local skill artifacts. Audits remain read-only; never repair product code, tests, CI or infrastructure.

Execution contract: The checklist defines completion. Track each item internally as PENDING, PROVEN with evidence, CLEARED with evidence its condition is absent, or UNPROVEN with a gap; reading, delegation, tool failure, a zero exit status, or a self-reported success is not proof; only the observed outcome is. Reconcile after each section. Before returning, resolve all PENDING, count only PROVEN and CLEARED, and apply verdict and approval rules to every gap. Preserve intent, scope, and existing authorization. Continue authorized work; ask only for consequential unresolved choices or required external approval. When no one can answer during the run, state the exact question and apply the skill's verdict for the remaining gap instead of waiting or guessing. Scale depth to material risk without skipping checks. Preserve dependency and safety order; otherwise choose an appropriate verification method. Accept equivalent user or repository evidence; no other skill, named artifact, or complete lifecycle is required. Preserve source requirement and decision IDs. Bind reused evidence to relevant source versions, dirty changes, configuration, and environment; invalidate only affected claims. Choose one primary workflow by the requested outcome, honoring explicit selection. Consume neighboring contracts and valid evidence without duplicating ownership or restarting a full audit. Split genuinely distinct deliverables by scope, not by mandatory skill invocation. Broad independent assessments belong to audit workflows; focused reviews may stay with the domain workflow but remain read-only. On continuation, reconcile task, authorization, current state, and unresolved evidence. For long work, return a compact continuation record or update an already authorized artifact; read-only skills do not persist it. Distinguish artifact readiness, verified behavior, and external-action authority. Prepare authorized work before required approval. If blocked by an instruction, cite its exact source and unresolved boundary; do not invent approval gates from caution. Before a new or changed technical decision, establish applicability and actual technology versions, then verify version-dependent assumptions against official documentation or specifications even when confident. Search and read primary pages for current external recommendations; search snippets are not evidence. Version-matched bundled documentation or still-valid prior evidence can suffice. Record the decision, applicable version, source and limitations. If evidence is unavailable, mark the affected conclusion UNVERIFIED and block only dependent work when essential. Prefer native mechanisms; add tools or abstractions only for an evidenced requirement. Reuse the existing project rule owner instead of creating parallel policy.

Tool Routing

NeedPreferred capabilityFallback
Actual stack and ownersInstructions, manifests, lockfiles, focused source/configuration reads and Git stateSupplied artifacts with explicit gaps; never infer versions from examples
External semanticsVersion-matched official documentation, specifications and maintained upstream sourceBundled official docs or valid prior evidence; mark unsupported conclusions UNVERIFIED
Authorized changesProject-native editor, configuration and maintained runtime toolingMinimal scoped patch; BLOCKED when no safe mutation capability exists
Observable verificationExisting checks and isolated runtime/consumer exercisesStatic evidence labeled as such; BLOCKED when essential behavior cannot be established

Boundaries

Architecture requirements own quality targets; architecture decisions own structural tradeoffs. This workflow translates those inputs into project technical guidance, not a second architecture baseline. Executable configuration owns exact values. Local procedures own steps, not copied policy. A readiness request does not authorize implementing every gap.

Checklist

1. Establish applicability

  • Inspect requested artifacts, Git state and existing instructions; preserve unrelated and interrupted work.
  • Identify project type, runtimes/versions, deployment topology, trust boundaries and operational constraints from executable sources.
  • Assess containers, configuration, authentication, authorization, errors, observability, mutation recovery and database lifecycle separately; record applicability and evidence for exclusions.
  • Read existing requirements and contracts; distinguish accepted current behavior, desired changes and unverified historical claims.
  • Locate each normative rule owner and executable owner of mutable values; resolve material conflicts without replacing business intent.

2. Build project guidance

  • Map applicable principles to native stack mechanisms and version-matched evidence; record accepted exceptions and revisit triggers.
  • Reuse existing documents; create an owner only when none exists. Link architecture targets and functional grants instead of copying them.
  • Keep always-loaded instructions limited to constraints and conditional routing. Use supported host imports or references to one canonical body; verify actual resolution in each targeted host.
  • Create a local skill only for an evidenced recurring procedure. Derive commands from the project; include inputs, target identity, side effects, verification, recovery and continuation.
  • Keep local proc
阅读完整来源 ↗