claude/skills/tech_text2ir/SKILL.md
技術記事/RFC/ADR/ポストモーテム等をTechArticleIR(Procedure/Design+Evidence+Discourse)に変換する。手順・意思決定・落とし穴を直交に抽出。「技術記事をIRに変換して」「TechIR作って」「RFCを構造化」「ADRをYAMLに」等で使用。ir-pipelineのtext2ir工程でも自動呼出。
npx skillsauth add lilpacy/dotfiles tech_text2irInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
あなたは「技術文書→IRコンパイラ」です。入力から TechArticleIR を生成してください。
ir_version: "techir.v0"
core:
metadata: { title: "", context: "", id: "" }
anchors:
- id: "A1"
kind: "paragraph"
path: "S1.P1"
quote: ""
entities:
- id: "E1"
type: "Component|API|Config|Metric|Env|Actor|Term"
surface: ["..."]
note: ""
facets:
discourse:
sections:
- id: "S1"
heading: ""
anchors: ["A1"]
procedure_design:
goals:
- id: "G1"
statement: ""
anchors: ["A2"]
decisions:
- id: "D1"
statement: ""
rationale: ""
anchors: ["A3"]
tags: ["ADR?","RFC?"]
tradeoffs:
- id: "T1"
pros: []
cons: []
anchors: ["A4"]
steps:
- id: "P1"
kind: "Step|Checklist|Migration|Runbook"
statement: ""
preconditions: ["E?","P?"]
anchors: ["A5"]
pitfalls:
- id: "F1"
statement: ""
mitigation: ""
anchors: ["A6"]
links:
- type: "requires|enables|blocks|alternative_to|mitigates|leads_to"
from: "P1"
to: "P2"
anchors: ["A5"]
confidence: 0.0
evidence:
items:
- id: "V1"
kind: "Benchmark|Incident|Spec|Doc|Observation|Experiment"
summary: ""
anchors: ["A7"]
links: ["D?","P?"]
measures:
- metric: ""
value: ""
delta: ""
links:
- type: "evidences|contradicts"
from: "V1"
to: "D1"
anchors: ["A7"]
confidence: 0.0
exceptions:
notes:
- id: "X1"
kind: "CustomerSpecific|EnvSpecific|StateSpecific|Legacy"
statement: ""
anchors: ["A8"]
risk: "low|med|high"
uncertainty:
notes:
- id: "U1"
kind: "Unverified|NeedsTest|EnvDependent|Assumption"
statement: ""
anchors: ["A9"]
severity: "low|med|high"
plan:
surface_templates:
- type: "howto|rfc|adr|postmortem|blog"
outline:
- role: "Problem"
include: ["S?","G?"]
- role: "Decision"
include: ["D?","T?","V?"]
- role: "Steps"
include: ["P?","F?","X?"]
validation:
invariants:
- "Every decision/step/pitfall/uncertainty has >=1 anchor"
- "Evidence links must cite anchors"
- "No duplicated decision text inside evidence/discourse; link by IDs"
data-ai
ユーザー向けの回答、要件整理、説明、計画、レビュー、仕様、要約を作成するときに使う。自然言語だけでは条件、状態、多重度、期間、時刻、境界値、制約、推論、計算が曖昧になりうる場合、表、デシジョンテーブル、Mermaid 図、計算式、境界値表で一意化する。推測で補った図表要素は必ず明示する。
tools
Cross-agent messaging via SQLite. Send messages between Claude Code, Codex, Gemini CLI, and other agents. No daemon, no network, no dependencies beyond bash and sqlite3.
development
Anti-AI-slop design skill for greenfield pages, audits, redesigns, and design extraction from URLs or screenshots. Use when the user asks to build a new app or landing page, wants to redesign something, invokes Hallmark by name, or uses audit/redesign/study.
development
Reviews animation and motion code against a high craft bar derived from Emil Kowalski's design engineering philosophy. Default to flagging; approval is earned.