DS

dualface/ste-zht

Developer tools

This page was translated automatically. Technical details may differ from the original. View the original

32 stars Quality 68 Trend 28MIT

Agent skill: Reports results to users in Traditional Chinese according to the ASD-STE100 principle

Overview

A task reporting specification skill for AI Agents. Drawing on the rigorous principles of the aviation and industrial industries, Agents are required to report work results in clear, rigorous, and unambiguous Traditional Chinese. Agents' daily reporting is often lengthy, inconsistent, or missing key verification evidence. After enabling ste-zht, Agents will report according to the controlled Traditional Chinese standard:. 💡 This skill only specifies how Agents express themselves when reporting and discussing with users, and does not affect project code, comments, commit information, or generated file content. Traditional Chinese version. For Simplified Chinese, please use the sister project ste-zh. Terminology is based on the Taiwan Traditional terminology list compiled by contributor laihenyi at ste-zh#1; This repository is an independent skill, not a machine translation for Simplified and Traditional Chinese. For the same bug fix task, the differences between common reports and normalized reports are as follows. The following excerpt is based on the sample result report. The source facts in the example are found in the file. "Alright, I first checked the login-related code and found that the problem probably lies with the token update logic. It seems to be caused by two requests updating simultaneously during concurrent processing. Then I modified auth/refresh.ts so that only one update request is sent at a time, while other requests wait for the result of this update. I also optimized some logs and similar files as well. Most tests have been completed, so there shouldn't be any issues. However, iOS can't verify it temporarily due to lack of environment. Do you want me to change the retry logic as well?" Completed: Code modification for login update logic. iOS side not verified. : Two requests to update tokens simultaneously. Later completed updates invalidate the tokens obtained first. : - auth/refresh.ts:42 — No check is made before updating to see if updates are in progress. - auth/refresh.ts: Only one update request is sent at the same time. Other requests await the result of this update. - auth/logger.ts: Deleted 3 duplicate logs. This change is outside the original scope. - Verified: npm test auth — 48 passed, 0 failed. Includes new concurrent update test auth/refresh.test.ts:88. - Unverified: iOS side — Locally no iOS build environment. - Whether auth/logger.ts changes are retained: 1. Revert (recommended) — This commit only includes login fixes. 2. Keep — Submit log cleaning and login repair together. - : The first sentence should be written according to the type of reply.

Overview (machine translated)

Language: Chinese Simplified | Traditional Chinese

A task reporting specification skill for AI Agents. Drawing on the rigorous ASD-STE100 (Simplified Technical English) principles from the aviation and industrial sectors, agents are required to report work results in clear, rigorous, and unambiguous Traditional Chinese.

Agents’ daily reporting is often lengthy, vague in assumptions, or missing key verification evidence. After enabling ste-zht, Agents will report according to controlled Traditional Chinese standards: Key information first, evidence-based speaking, fixed status, clear at a glance.

💡 > This skill only specifies how agents express themselves when reporting and discussing with users, and does not affect project code, comments, commit information, or generated file content.

Traditional Chinese version. For Simplified Chinese, please use the sister project ste-zh. Terminology is based on the Taiwan Traditional terminology list compiled by contributor laihenyi at ste-zh#1; This repository is an independent skill, not a machine translation for Simplified and Traditional Chinese.


Why do you need ste-zht?

For the same bug fix task, the differences between common reports and normalized reports are as follows. The following excerpt is based on the sample result report. The source facts in the example are found in the file.

Common AI reports (lengthy, speculative, non-empirical):

“Okay, I first checked the login code and found that the problem probably lies with the token update logic. It’s probably caused by two requests updating simultaneously during concurrent processing. Then I modified auth/refresh.ts so that only one update request is sent at a time, while other requests wait for the results of this update. I also optimized some logs and such. Most tests have been completed, so there shouldn’t be any issues, but iOS can’t verify it temporarily due to lack of environment. Do you want me to change the retry logic as well?”

ste-zht Standardized reporting (conclusions first, evidence-based, facilitating decision-making):

Completed: Code modification for login update logic. iOS side not verified.

Cause: Two requests update tokens simultaneously. Updates completed later invalidate the tokens obtained first.

Evidence:

  • auth/refresh.ts:42 — Do not check if an update is ongoing before the update.

Changes

  • auth/refresh.ts: Only one update request is sent at the same time. Other requests await the result of this update.
  • auth/logger.ts: Deleted 3 duplicate logs. This change is outside the original scope.

Verification

  • Verified: npm test auth — 48 passed, 0 failed. Includes the addition of concurrent update test auth/refresh.test.ts:88.
  • Unverified: iOS side — The device does not have an iOS build environment.

To be confirmed

  • Whether to keep changes to auth/logger.ts:
  1. Retraction (Recommendation) — This submission only includes login fixes.

This page was translated automatically. Technical details may differ from the original.

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add dualface/ste-zht
New to MCP? Start with the mcpskill.org guide →