ワンクリックでManusで任意のスキルを実行

mecris-bonsai

スター4

フォーク1

更新日2026年5月9日 23:49

Maintain project health through intentional pruning, context hygiene, and rigorous validation of multilingual accountability features. Trigger with /bonsai-orient

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

Manusで実行

ソース

kingdonb

kingdonb/mecris

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

Manusで実行

Mecris Bonsai: The Willow Kitten Path

I ensure the Mecris ecosystem grows intentionally, maintaining strict context hygiene and verifying the behavioral integrity of our multilingual accountability system.

When I Activate

"bonsai", "shaping", "pruning"
"context hygiene", "prevent context burn"
"validate android", "test coaching"
"Arabic reminders", "Greek script"

Core Capabilities

1. Context Hygiene (Anti-Burn)

I enforce strict output limitations to prevent runaway token consumption during agentic loops.

Rule: NEVER use -v in pytest commands.
Enforcement: Always append -q --tb=short to test executions.

2. Backlog Curation (Bonsai Shaping)

I maintain the "Willow Kitten" philosophy: Depth over Breadth.

Pruning: Identify and move non-essential features (Chrome bookmarks, bidding markets) to attic/DORMANT_BACKLOG.md.
Integration: Focus energy on the "Trunk" (Goal 1: Nagging, Goal 2: Knowledge Base).

3. Multilingual Verification

I ensure that accountability messages for Arabic and Greek goals are culturally and linguistically resonant.

Script Validation: Verify that messages containing Arabic/Greek logic actually use the correct Unicode ranges (0x0600-0x06FF for Arabic, 0x0370-0x03FF for Greek).

4. Android Build Validation

I leverage the environment's bundled tools to verify frontend integrity.

Mandate: Use the Android Studio bundled JDK for all Gradle tasks.
Command: export PATH="/Applications/Android Studio.app/Contents/jbr/Contents/Home/bin:$PATH" && ./gradlew test

Slash Command

`/bonsai-orient`

Runs a full health audit of the workspace:

Audit Context: Check TDG.md for verbose flags.
Audit Backlog: Verify ACTIVE_BACKLOG.md is consolidated.
Verify Build: Run Android unit tests and Python coaching tests.
Verify Multilingual: Run script validation checks on the latest coaching insights.

Script Verification: Before executing, verify the script integrity:

sha256sum .github/skills/mecris-bonsai/scripts/validate.sh
# Expected: e0be6e14eddb96f9b6680c36a1cc026e79dfebb342b901b526577c7029f71d49

Expected Failure Modes

Failure Mode	Symptoms	Workaround
JDK Not Found	Android build fails with "Java not found"	Export path to Android Studio bundled JBR
Context Burn	API credits depleted rapidly	Check for verbose logs or massive file reads
Script Corruption	Non-Latin characters mangled in terminal	Verify locale and UTF-8 encoding

MCP Server Integration

Use the get_narrator_context tool to assess the "System Pulse" before performing any pruning, ensuring the agent is aware of the current strategic intent.

このリポジトリの他の Skills

同じリポジトリ

beeminder-audit

kingdonb/mecris

Procedural guidance on safely auditing, modifying, and cleaning up Beeminder goal data (including odometer resets, @TARE mechanics, and timezone protections) without risking accidental derailment. Triggers on keywords like beeminder audit, beeminder cleanup, delete datapoints, tare.

2026-06-304

mecris-archive

kingdonb/mecris

Serialize session state after work is complete. Closes the plan issue, updates NEXT_SESSION.md, appends to SESSION_LOG, and commits. Handles partial completion honestly. The on-save arm. Trigger with /mecris-archive

2026-04-044

mecris-plan

kingdonb/mecris

Write a minimal spec as a GitHub issue before acting. Takes the recommended action from mecris-orient and records intent, rationale, and validation criteria. This is the red in red-green-archive. Trigger with /mecris-plan

2026-04-044

mecris-rebase

kingdonb/mecris

Safely resolves git conflicts and ensures successful pushes during autonomous sessions. Use when git pull --rebase fails.

2026-04-024

talktype-mirror

kingdonb/mecris

Manage the synchronization and IP separation between the public talktype-repo and the internal Urmanac mirror. Use when developing features across public and internal boundaries, ensuring Urmanac-specific context remains isolated while generic improvements are shared.

2026-03-304

tdg-test-driven-generation

kingdonb/mecris

Test-Driven Generation uses TDD (Test-Driven Development) techniques to generate tests and code in Red-Green-Refactor loops. Key words to detect are tdg, TDG.

2026-03-274

name	mecris-bonsai
description	Maintain project health through intentional pruning, context hygiene, and rigorous validation of multilingual accountability features. Trigger with /bonsai-orient
allowed-tools	["read_file","run_shell_command","grep_search","write_file","replace"]

Mecris Bonsai: The Willow Kitten Path

I ensure the Mecris ecosystem grows intentionally, maintaining strict context hygiene and verifying the behavioral integrity of our multilingual accountability system.

When I Activate

"bonsai", "shaping", "pruning"
"context hygiene", "prevent context burn"
"validate android", "test coaching"
"Arabic reminders", "Greek script"

Core Capabilities

1. Context Hygiene (Anti-Burn)

I enforce strict output limitations to prevent runaway token consumption during agentic loops.

Rule: NEVER use -v in pytest commands.
Enforcement: Always append -q --tb=short to test executions.

2. Backlog Curation (Bonsai Shaping)

I maintain the "Willow Kitten" philosophy: Depth over Breadth.

Pruning: Identify and move non-essential features (Chrome bookmarks, bidding markets) to attic/DORMANT_BACKLOG.md.
Integration: Focus energy on the "Trunk" (Goal 1: Nagging, Goal 2: Knowledge Base).

3. Multilingual Verification

I ensure that accountability messages for Arabic and Greek goals are culturally and linguistically resonant.

Script Validation: Verify that messages containing Arabic/Greek logic actually use the correct Unicode ranges (0x0600-0x06FF for Arabic, 0x0370-0x03FF for Greek).

4. Android Build Validation

I leverage the environment's bundled tools to verify frontend integrity.

Mandate: Use the Android Studio bundled JDK for all Gradle tasks.
Command: export PATH="/Applications/Android Studio.app/Contents/jbr/Contents/Home/bin:$PATH" && ./gradlew test

Slash Command

`/bonsai-orient`

Runs a full health audit of the workspace:

Audit Context: Check TDG.md for verbose flags.
Audit Backlog: Verify ACTIVE_BACKLOG.md is consolidated.
Verify Build: Run Android unit tests and Python coaching tests.
Verify Multilingual: Run script validation checks on the latest coaching insights.

Script Verification: Before executing, verify the script integrity:

sha256sum .github/skills/mecris-bonsai/scripts/validate.sh
# Expected: e0be6e14eddb96f9b6680c36a1cc026e79dfebb342b901b526577c7029f71d49

Expected Failure Modes

Failure Mode	Symptoms	Workaround
JDK Not Found	Android build fails with "Java not found"	Export path to Android Studio bundled JBR
Context Burn	API credits depleted rapidly	Check for verbose logs or massive file reads
Script Corruption	Non-Latin characters mangled in terminal	Verify locale and UTF-8 encoding

MCP Server Integration

Use the get_narrator_context tool to assess the "System Pulse" before performing any pruning, ensuring the agent is aware of the current strategic intent.