CtrlK
BlogDocsLog inGet started
Tessl Logo

pua-ja

お前のAIを詰めろ。日本企業の詰め文化と体系的デバッグ方法論で全ての手段を尽くさせる。トリガー条件:(1) タスク失敗2回以上または同じアプローチの微調整ループ; (2)「解決できません」と言おうとする・手動対応を推奨・未検証で環境を原因帰属; (3) 受け身——検索しない・ソースを読まない・指示待ち; (4) ユーザーの不満:'もっと頑張れ'、'なんでまた失敗したの'、'なんとかしろ'。全タスクタイプ適用。初回失敗や既知修正の実行中はトリガーしない。

69

Quality

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

The canonical home for this skill is pua-ja in tanweai/pua

SKILL.md
Quality
Evals
Security

Quality

Content

77%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-sequenced methodology with concrete checklists and validation gates, but it is verbose with redundant motivational rhetoric and ships as one large monolithic file rather than splitting its reference material into bundle files.

Suggestions

Move the 大企業詰め拡張パック flavor library and the 言い訳封殺テーブル into a references/ file (e.g. FLAVORS.md) and link to them from a concise overview, recovering both conciseness and progressive_disclosure points.

Deduplicate the repeated 「証拠を出せ / 口で検証するな」 message that appears across 能動性の鞭撻フレーズ, the 検証型 checklist, and multiple flavor packs into a single canonical statement.

Trim the rhetorical framing in the 三つの鉄則 and Step introductions to assume Claude's competence and reduce token cost without losing the actionable sub-steps.

DimensionReasoningScore

Conciseness

The body is ~375 lines / 37KB with substantial redundancy — the 言い訳封殺テーブル, 能動性の鞭撻フレーズ, and the ten flavor packs repeatedly restate the same 「verify, don't just claim done」 message — so it could be tightened considerably despite being mostly purposeful.

2 / 3

Actionability

It gives concrete executable guidance: a 5-step methodology with specific actions (「前後50行」「公式ドキュメント + Issues を検索」), a 7-item checklist with per-task-type equivalents, escalation-forced actions, and a copy-paste [PUA-REPORT] format.

3 / 3

Workflow Clarity

Multi-step processes are clearly sequenced (Steps 1–5, escalation L1–L4 by failure count, selector rounds 第一~最終ラウンド) with explicit validation gates such as 「次元1-4が完了するまでユーザーへの質問は禁止」 and the 7-item checklist as a completion gate.

3 / 3

Progressive Disclosure

No bundle files exist (references/scripts/assets absent) and the skill is a large monolithic SKILL.md; the flavor-pack library and excuse table are content that could be split into separate reference files, so it is structured but not appropriately split, capping this dimension below 3.

2 / 3

Total

10

/

12

Passed

Description

90%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A specific, well-triggered description that clearly states both capability and activation conditions with natural user-facing trigger phrases. The only weakness is the second-person imperative voice, which the rubric penalizes on specificity.

Suggestions

Rewrite the opening in third person (e.g. 「AIを諦めさせず全ての手段を尽くさせる」) instead of the second-person imperative 「お前のAIを詰めろ」 to recover the specificity point lost to voice.

DimensionReasoningScore

Specificity

It lists multiple concrete trigger conditions and a named methodology (「体系的デバッグ方法論」「詰め文化」), which would anchor at 3, but it opens with the second-person imperative 「お前のAIを詰めろ」 and the rubric penalizes non-third-person voice by −1 on this dimension.

2 / 3

Completeness

It explicitly answers both what (「全ての手段を尽くさせる」 via 詰め文化 + 体系的デバッグ方法論) and when (four numbered trigger conditions plus the explicit exclusion 「初回失敗や既知修正の実行中はトリガーしない」).

3 / 3

Trigger Term Quality

It includes natural user phrases a person would actually say — 「もっと頑張れ」「なんでまた失敗したの」「なんとかしろ」 — giving good coverage of real trigger terms rather than technical jargon.

3 / 3

Distinctiveness Conflict Risk

The PUA/詰め methodology is a distinct niche, and the specific trigger set plus the negative trigger (no fire on first failure or known-fix execution) make it unlikely to fire for the wrong skill.

3 / 3

Total

11

/

12

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
tanweai/pua
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.