lilmgenius/paperthin

sip

After you create or change an artifact or skill, taste-test it with our own skills instead of trusting your in-session judgment.

ソースを見る

使い方

  1. インストールコマンドをコピーし、自分の環境へ導入する前にソースを確認します。
  2. 必要なツール、認証情報、対応環境を原文で確認します。
  3. 入力と成果物を指定し、出力を資料と照合します。
npx skills add https://github.com/lilmgenius/paperthin --skill sip

開始前の確認

  • 導入にはモデル契約、API枠、ファイルへのアクセス許可は含まれません。
  • スター数や導入数は活動の指標であり、独立した品質・安全性試験ではありません。
リポジトリの原文

見出し、例、コード、表、リンク、参照画像を含む原文を表示しています。

Taste-test what you just made with this suite's own clarity and truth checks before you call it done.

Goal

Your judgment of a thing is most biased right after you make it. sip runs this suite's own clarity and truth checks on an artifact or skill you just created or changed, so its verification does not rest on that judgment. It fires on your own output, immediately after you produce it, when you are about to call it done, commit it, or hand it off. It exists because a written reminder to verify does not reliably fire in a fresh session; the skill is the trigger.

Workflow

Run the steps in this order.

  1. Cold read. Call the Skill tool with "shower" for a fresh reader's comprehension and handoff check.
  2. Truth. Call the Skill tool with "factchk" for claims grounded in reality, and call it separately with "mandela" for an eval, metric, or experiment. Skip this step when the artifact has neither kind of content.
  3. Consistency. Call the Skill tool with "ssotize" to audit the repository for anything the change duplicated or contradicted. Execute its consolidation plan only after approval.
  4. Portability. Call the Skill tool with "detool" only when the artifact claims portability, tool-neutrality, stack-agnostic durability, or cross-agent reuse. Skip it for provenance, operational notes, runbooks aimed at a specific tool, and anything making no such claim.
  5. Tidy. Call the Skill tool with "re0" on the changed docs so they read as a clean v0 with no traces of patching.
  6. Apply and hand over. You return the findings to the author session, which applies each one or defers it with a reason, then hands the artifact over.

Rules

  • Use the named skills; never reimplement their checks inline. shower covers clarity, factchk and mandela truth, ssotize single source of truth, detool portability claims, and re0 cleanup.
  • sip orchestrates only: findings go back to the author session, which makes the fixes.
  • Skip any check that plainly does not apply, and any whose skill is not installed. factchk and mandela need a claim or an eval; detool needs one of the four portability claims in step 4. A one-line prose tweak may need only the consistency check.
  • State every skip with its reason.
  • Stop at the artifact: never touch version control and never commit.
  • Call only model-invoked skills. A user-invoked skill is one its host keeps out of the model's reach (a disable-model-invocation frontmatter flag, for one); only a human can fire it, so sip never calls one.
  • The only approval gate is ssotize's consolidation plan in step 3.

Verification

Before handing over, confirm:

  1. The relevant skills actually ran on the change; an informal look does not count.
  2. The author session applied each finding, or deferred it on purpose with a stated reason.
  3. Either the artifact changed as a result (a diff exists), or every check that ran found nothing and every skipped skill has a stated reason; in that case the artifact is handed over unchanged with that list.
同じリポジトリから

関連する Skills

すべての Skills
lilmgenius
コミュニティ

re0-order

Move items only, reword nothing; realign a listing whose order has gone arbitrary (a list, table, catalog, sections, an enum) into a logical sequence under one stated principle. Nothing is added or removed. Use when the user names a listing, or a scope holding one, whose order no longer helps a reader follow the set, or when a pipeline the user invoked passes that scope on. Acts only inside that scope.

1時間の導入数
5
GitHub Stars
1463
更新日
11月15日
lilmgenius
コミュニティ

readback

Verify the model's understanding of a user's instruction before spending non-trivial work. Use when a request is long, bundled, high-stakes, hard to undo, or has ambiguous scope or referents such as this, that, it, the other one, whatever is cleaner, or whichever order makes sense. Restate internally, cross-check against available context, proceed silently when resolved, and surface only a genuine surviving fork.

1時間の導入数
5
GitHub Stars
1463
更新日
11月15日
lilmgenius
コミュニティ

whychk

Right after you pick an option, have a fresh outsider find the sharpest gap in the choice, then explain that gap in your own words, one gap at a time.

1時間の導入数
5
GitHub Stars
1463
更新日
11月15日
lilmgenius
コミュニティ

aim

When data lands with a thin ask (files, a repo, a paste, a chat scrollback), read it and propose what the user is trying to do with it for them to confirm. Use the moment material arrives without a fully-formed ask, or when the ask reads thinner than the data suggests.

1時間の導入数
4
GitHub Stars
1463
更新日
11月15日