content-calibrator

内容校准预测循环——打分+盲预测合一 → 发布 → 记录 → T+3d 复盘 → 进化 rubric。本技能负责打分+预测（blind sub-agent + score-only.sh + commit-prediction.sh + 阈值门）与校准闭环；发布记录与数据采集由 published-track 统一管理。

Quality

61%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./crews/main/skills/content-calibrator/SKILL.md

Quality

Content

63%Weight 40%Scale 1-5

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

The body is highly actionable with clearly sequenced, validated workflows and a well-organized script bundle. Its weaknesses are structural repetition across the flow and Scripts sections, a missing referenced script, and no progressive disclosure of the heavy conceptual content into reference files.

Suggestions

Remove the duplicate script descriptions: keep executable examples in one place (either the flow or the Scripts section) and cross-reference from the other.

Add the missing query-retro-pending.sh to the scripts bundle, or replace references with an existing script (e.g. query-metrics.sh).

Split the detailed rubric formula, bump-detection algorithm, and retro logic into reference files (e.g. RUBRIC.md, BUMP.md) referenced one level deep from SKILL.md.

Dimension	Reasoning	Score
Conciseness	Operational and mostly free of basic-concept padding, but noticeably repetitive: score-only.sh, commit-prediction.sh, detect-bump-signals.sh, and validate-rubric.sh are each fully described twice (once in the flow sections, again in the Scripts section), and the dispatch strategy is restated multiple times.	3 / 5
Actionability	Provides copy-paste-ready commands with full arguments (score-only.sh, commit-prediction.sh, validate-rubric.sh) covering common cases; docked one point because query-retro-pending.sh is referenced in the Retro flow but is absent from the scripts bundle.	4 / 5
Workflow Clarity	Multi-step flows (1A steps 1-6, Retro 3a→3b, Bump 1-5) are clearly sequenced with explicit validation checkpoints and feedback loops (阈值门最多2轮, validate-rubric 最多3轮, batch-then-aggregate ordering); minor gaps include the missing query-retro-pending.sh and an overwrite-without-verify step in commit-prediction.sh.	4 / 5
Progressive Disclosure	Well-sectioned with a route table and file-structure diagram, and the scripts bundle is referenced one level deep, but all conceptual content (rubric dims, full flow logic, bump/retro algorithms) lives inline in a ~487-line monolith with no reference .md files to split out heavier material.	3 / 5
	Total	14 / 20 Passed

Description

58%Weight 40%Scale 1-5

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

The description is specific and well-bounded, naming concrete actions, tools, and an explicit division of labor with published-track. Its main weakness is the absence of an explicit 'Use when…' trigger clause and reliance on jargon over natural user phrases.

Suggestions

Append an explicit trigger clause, e.g. 'Use when scoring/previewing a draft before publishing, doing a T+3d retrospective, or evolving the content rubric.'

Add natural-language synonyms users might actually say (e.g. '给这篇打分', '复盘', '升级评分公式') alongside the technical terms.

Mention the supported platforms or content forms briefly to further reduce overlap with published-track.

Dimension	Reasoning	Score
Specificity	Names several concrete actions and tools—'打分+盲预测合一', 'T+3d 复盘', '进化 rubric', '阈值门' plus specific scripts 'score-only.sh + commit-prediction.sh'—giving comprehensive coverage of its niche with only minor gaps.	4 / 5
Completeness	Clearly states what the skill does ('负责打分+预测…与校准闭环') but provides no 'Use when…' trigger clause, so completeness is capped at 3 per the missing-trigger-guidance rule.	3 / 5
Trigger Term Quality	Contains some relevant keywords ('打分', '预测', '复盘', 'rubric') but leans on technical jargon ('盲预测', '阈值门', '校准闭环') and lacks common synonyms or natural user phrasings.	3 / 5
Distinctiveness Conflict Risk	Has a clear niche (content calibration/prediction loop) and explicitly carves out a boundary ('发布记录与数据采集由 published-track 统一管理'), leaving only minor overlap risk with the closely related published-track skill.	4 / 5
	Total	14 / 20 Passed

Validation

87%

Warnings & errors only

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

Validation — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning

	Total	14 / 16 Passed

Repository: TeamWiseFlow/xiaobei
Path: crews/main/skills/content-calibrator/SKILL.md
Commit: fdc03d6

Reviewed: 13 days ago

Table of Contents

Discovery Implementation Validation

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.