mirror of
https://github.com/kochetkov-ma/claude-brewcode.git
synced 2026-09-14 20:16:41 +08:00
v3.7.7: compress skill/agent descriptions, trim think-short profiles, enforce description budget defaults
This commit is contained in:
kochetkov-ma
@@ -6,13 +6,13 @@
|
||||
},
|
||||
"metadata": {
|
||||
"description": "Claude Code plugin suite: brewcode for infinite task execution, brewdoc for documentation tools, brewtools for text utilities, brewui for UI/visual/creative tools",
|
||||
"version": "3.7.6"
|
||||
"version": "3.7.7"
|
||||
},
|
||||
"plugins": [
|
||||
{
|
||||
"name": "brewcode",
|
||||
"description": "Brewcode - full-featured development platform for Claude Code: infinite focus tasks, prompt optimization, skill/agent creation, quorum reviews, rules management",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"category": "productivity",
|
||||
"keywords": [
|
||||
"brewcode",
|
||||
@@ -46,7 +46,7 @@
|
||||
{
|
||||
"name": "brewdoc",
|
||||
"description": "Brewdoc - Claude Code documentation tools: auto-sync for skills/agents/rules, my-claude installation docs, memory optimization, md-to-pdf conversion",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"category": "productivity",
|
||||
"keywords": [
|
||||
"brewdoc",
|
||||
@@ -74,7 +74,7 @@
|
||||
{
|
||||
"name": "brewtools",
|
||||
"description": "Brewtools - universal utilities for Claude Code: text optimization, humanization, secrets scanning",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"category": "productivity",
|
||||
"keywords": [
|
||||
"brewtools",
|
||||
@@ -102,7 +102,7 @@
|
||||
{
|
||||
"name": "brewui",
|
||||
"description": "Brewui - UI/visual/creative tools for Claude Code: AI image generation, design-to-code conversion",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"category": "productivity",
|
||||
"keywords": [
|
||||
"brewui",
|
||||
|
||||
@@ -2,6 +2,33 @@
|
||||
|
||||
---
|
||||
|
||||
## v3.7.7 (2026-04-22)
|
||||
|
||||
> Docs: [brewtools:think-short](https://doc-claude.brewcode.app/brewtools/skills/think-short/) | [brewtools:agent-toggle](https://doc-claude.brewcode.app/brewtools/skills/agent-toggle/) | [brewtools:skill-toggle](https://doc-claude.brewcode.app/brewtools/skills/skill-toggle/) | [brewui:glm-design-to-code](https://doc-claude.brewcode.app/brewui/skills/glm-design-to-code/) | [brewui:image-gen](https://doc-claude.brewcode.app/brewui/skills/image-gen/) | [brewtools:text-optimizer](https://doc-claude.brewcode.app/brewtools/agents/text-optimizer/) | [brewui:glm-zai-specialist](https://doc-claude.brewcode.app/brewui/agents/glm-zai-specialist/) | [brewcode:hook-creator](https://doc-claude.brewcode.app/brewcode/agents/hook-creator/) | [brewcode:skill-creator](https://doc-claude.brewcode.app/brewcode/agents/skill-creator/) | [brewcode:agent-creator](https://doc-claude.brewcode.app/brewcode/agents/agent-creator/) | [brewcode:tester](https://doc-claude.brewcode.app/brewcode/agents/tester/) | [brewcode:developer](https://doc-claude.brewcode.app/brewcode/agents/developer/)
|
||||
|
||||
### brewtools
|
||||
#### Changed
|
||||
- **think-short:** profile directives compressed with zero semantic loss. `light` ~35 -> ~20 tok, `medium` ~70 -> ~35 tok, `aggressive` ~200 -> ~120 tok. Removed redundant phrase examples, README-level justifications after dashes, duplicated directives. Critical `User instructions always override these rules` retained in aggressive. All profiles now ASCII-only (em-dash -> hyphen), matching the profile's own rule.
|
||||
- **think-short (SKILL.md):** description frontmatter compressed ~480 -> ~60 tokens. English-only triggers. Lead sentence <=160 chars.
|
||||
- **agent-toggle / skill-toggle (SKILL.md):** descriptions compressed ~255 -> ~45 tokens each. English-only triggers.
|
||||
- **text-optimizer (agent):** description tuned to ~170 tokens - kept 2 examples since it is a frequently invoked agent.
|
||||
- **session-start hook:** dropped redundant `brewtools: active | session: X` line from `additionalContext` (retained in `systemMessage` UI log). Saves ~30 tokens per session.
|
||||
|
||||
### brewui
|
||||
#### Changed
|
||||
- **glm-design-to-code / image-gen (SKILL.md):** descriptions compressed ~55% each. English-only triggers.
|
||||
- **glm-zai-specialist (agent):** description compressed ~247 -> ~50 tokens.
|
||||
- **session-start hook:** same cleanup as brewtools - removed `active | session:` line from `additionalContext`.
|
||||
|
||||
### brewcode
|
||||
#### Changed
|
||||
- **hook-creator / skill-creator / agent-creator (agents):** descriptions tuned to ~180 tokens each with 2 high-signal examples - these agents are frequently invoked so examples retained intentionally.
|
||||
- **tester / developer (agents):** descriptions compressed ~35%, example blocks removed (triggers already convey intent).
|
||||
- **skill-creator / agent-creator:** new "Description Budget" section enforces <=100 tok (skills) / <=150 tok (agents) default, English-only triggers, <=1 example for future creations.
|
||||
- **skills/skills/SKILL.md / skills/agents/SKILL.md:** creation flows now carry the same description-budget rule inline.
|
||||
|
||||
---
|
||||
|
||||
## v3.7.6 (2026-04-21)
|
||||
|
||||
> Docs: [brewtools:think-short](https://doc-claude.brewcode.app/brewtools/skills/think-short/)
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "brewcode",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"description": "Brewcode - full-featured development platform for Claude Code: infinite focus tasks, prompt optimization, skill/agent creation, quorum reviews, rules management",
|
||||
"author": {
|
||||
"name": "Maksim Kochetkov",
|
||||
|
||||
@@ -1,16 +1,16 @@
|
||||
---
|
||||
name: agent-creator
|
||||
description: |
|
||||
Creates, improves, and analyzes Claude Code agents. Triggers: "create agent", "new agent", "agent doesn't trigger", "improve agent", "fix agent description".
|
||||
Creates, improves, and analyzes Claude Code agents - tunes frontmatter, tools, model, permission mode, and system prompt for reliable delegation. Triggers: create agent, new agent, agent doesn't trigger, improve agent, fix agent description, refactor agent, scaffold agent.
|
||||
|
||||
<example>
|
||||
user: "Create an agent for code review"
|
||||
<commentary>Explicit agent creation request triggers this agent</commentary>
|
||||
<commentary>Scaffolds agent .md with delegation-ready description.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "My reviewer agent doesn't trigger reliably"
|
||||
<commentary>Agent improvement/debugging triggers this agent</commentary>
|
||||
<commentary>Diagnoses description, triggers, and model selection.</commentary>
|
||||
</example>
|
||||
model: opus
|
||||
color: cyan
|
||||
@@ -21,6 +21,18 @@ tools: Read, Write, Edit, Glob, Grep, Task, Skill, WebFetch, WebSearch, AskUserQ
|
||||
|
||||
Creates Claude Code agents following Anthropic best practices.
|
||||
|
||||
## Description Budget (DEFAULT)
|
||||
|
||||
| Constraint | Value |
|
||||
|------------|-------|
|
||||
| Total | <= 150 tokens (~600 chars) |
|
||||
| Lead sentence | <= 160 chars, plain EN prose |
|
||||
| Triggers | comma-list, EN only, 3-7 keywords |
|
||||
| Examples | at most 1, commentary <= 15 words |
|
||||
| Language | English only in frontmatter |
|
||||
|
||||
> Exceed only if user explicitly asks. For often-invoked agents user may allow up to ~200 tokens with 1-2 examples.
|
||||
|
||||
## Agent File Format
|
||||
|
||||
```markdown
|
||||
|
||||
@@ -1,12 +1,6 @@
|
||||
---
|
||||
name: developer
|
||||
description: |
|
||||
Implements features, writes code, fixes bugs, refactors - full-stack development. Triggers: "implement", "write code", "fix bug", "add feature", "refactor", "build".
|
||||
|
||||
<example>
|
||||
user: "Add pagination to the user list API"
|
||||
<commentary>Feature implementation triggers developer, not reviewer or tester</commentary>
|
||||
</example>
|
||||
description: "Implements features, writes code, fixes bugs, refactors - full-stack development. Triggers: implement, write code, fix bug, add feature, refactor, build."
|
||||
model: opus
|
||||
tools: Read, Write, Edit, Glob, Grep, Bash, Task, NotebookEdit, WebFetch, WebSearch
|
||||
auto-sync: true
|
||||
|
||||
@@ -1,16 +1,16 @@
|
||||
---
|
||||
name: hook-creator
|
||||
description: |
|
||||
Creates, debugs, and analyzes Claude Code hooks (bash/JS lifecycle event handlers). Triggers: "create hook", "PreToolUse hook", "Stop hook", "hook doesn't work", "debug hook".
|
||||
Creates, debugs, and analyzes Claude Code hooks - bash/JS lifecycle event handlers (SessionStart, PreToolUse, PostToolUse, Stop, PreCompact, etc.). Triggers: create hook, PreToolUse hook, SessionStart hook, Stop hook, hook doesn't work, debug hook, lifecycle event, hooks.json.
|
||||
|
||||
<example>
|
||||
user: "Create a PreToolUse hook to validate Bash commands"
|
||||
<commentary>Explicit hook creation request triggers this agent</commentary>
|
||||
<commentary>New hook with matcher and JSON schema.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "My Stop hook blocks even when task is complete"
|
||||
<commentary>Hook debugging triggers this agent</commentary>
|
||||
<commentary>Debugging decision: block logic and exit codes.</commentary>
|
||||
</example>
|
||||
model: opus
|
||||
color: yellow
|
||||
|
||||
@@ -1,16 +1,16 @@
|
||||
---
|
||||
name: skill-creator
|
||||
description: |
|
||||
Creates, improves, and analyzes Claude Code skills (SKILL.md files). Triggers: "create skill", "new skill", "skill doesn't invoke", "improve skill", "fix skill trigger".
|
||||
Creates, improves, and analyzes Claude Code skills (SKILL.md files) with optimized triggers, frontmatter, and progressive disclosure. Triggers: create skill, new skill, skill doesn't invoke, improve skill, fix skill trigger, skill activation, SKILL.md.
|
||||
|
||||
<example>
|
||||
user: "Create a skill for database migrations"
|
||||
<commentary>Explicit skill creation request triggers this agent</commentary>
|
||||
<commentary>Scaffolds SKILL.md with activation-optimized description.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "Claude isn't picking up my skill automatically"
|
||||
<commentary>Skill invocation issues trigger this agent</commentary>
|
||||
<commentary>Diagnoses frontmatter, tunes triggers, boosts activation rate.</commentary>
|
||||
</example>
|
||||
model: opus
|
||||
color: green
|
||||
@@ -28,6 +28,18 @@ Experienced developers: skip explanations, move faster. Watch for context cues.
|
||||
|
||||
> Skills replace Commands. `.claude/commands/review.md` and `.claude/skills/review/SKILL.md` both create `/review`. Commands are legacy — create Skills.
|
||||
|
||||
## Description Budget (DEFAULT)
|
||||
|
||||
| Constraint | Value |
|
||||
|------------|-------|
|
||||
| Total | <= 100 tokens (~400 chars) |
|
||||
| Lead sentence | <= 160 chars, plain EN prose |
|
||||
| Triggers | comma-list, EN only, 3-6 keywords |
|
||||
| Examples | at most 1, commentary <= 15 words |
|
||||
| Language | English only in frontmatter (RU/other only in README) |
|
||||
|
||||
> Exceed only if user explicitly asks. For often-invoked skills user may allow up to ~200 tokens with 1-2 examples.
|
||||
|
||||
## ⚠️ Activation Reality
|
||||
|
||||
**Skills auto-activate only 20-50% of the time.** Known issue ([#10768](https://github.com/anthropics/claude-code/issues/10768), [#15136](https://github.com/anthropics/claude-code/issues/15136) — both closed NOT PLANNED).
|
||||
|
||||
@@ -1,17 +1,6 @@
|
||||
---
|
||||
name: tester
|
||||
description: |
|
||||
Runs tests, analyzes failures, debugs flaky tests - SDET/QA specialist. Triggers: "run tests", "tests failing", "flaky test", "debug test", "test coverage".
|
||||
|
||||
<example>
|
||||
user: "Run the tests, several are failing after my refactor"
|
||||
<commentary>Test execution and failure analysis triggers tester, not developer</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "This test passes sometimes and fails other times"
|
||||
<commentary>Flaky test debugging triggers tester</commentary>
|
||||
</example>
|
||||
description: "Runs tests, analyzes failures, debugs flaky tests - SDET/QA specialist. Triggers: run tests, tests failing, flaky test, debug test, test coverage."
|
||||
model: sonnet
|
||||
tools: Read, Write, Edit, Glob, Grep, Bash, Task
|
||||
---
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "claude-plugin-brewcode",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"description": "Infinite task execution with automatic handoff for Claude Code",
|
||||
"keywords": [
|
||||
"claude-code",
|
||||
@@ -36,6 +36,6 @@
|
||||
},
|
||||
"claude-plugin": {
|
||||
"name": "brewcode",
|
||||
"version": "3.7.6"
|
||||
"version": "3.7.7"
|
||||
}
|
||||
}
|
||||
|
||||
@@ -56,6 +56,10 @@ Agent management - create and improve Claude Code agents.
|
||||
|
||||
## Mode: create
|
||||
|
||||
### Description Budget (DEFAULT)
|
||||
|
||||
Frontmatter `description`: <= 150 tokens (~600 chars). Lead sentence <= 160 chars, plain EN. Triggers: comma-list, EN only. At most 1 example. No RU/other languages unless user explicitly asks. For often-invoked agents user may allow up to ~200 tokens.
|
||||
|
||||
### Step 1: AskUserQuestion (batch -- all questions in ONE call)
|
||||
|
||||
Ask all 3 questions in a single AskUserQuestion:
|
||||
|
||||
@@ -50,6 +50,10 @@ bash "${CLAUDE_SKILL_DIR}/scripts/list-skills.sh" && echo "✅ list" || echo "
|
||||
|
||||
Both `create` and `up` follow Phases 0-6. Differences noted per phase.
|
||||
|
||||
### Description Budget (DEFAULT)
|
||||
|
||||
Frontmatter `description`: <= 100 tokens (~400 chars). Lead sentence <= 160 chars, plain EN. Triggers: comma-list, EN only. At most 1 example. No RU/other languages unless user explicitly asks.
|
||||
|
||||
### Prerequisite (up only): Resolve Target
|
||||
|
||||
**EXECUTE** using Bash tool:
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "brewdoc",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"description": "Brewdoc - Claude Code documentation tools: auto-sync for skills/agents/rules, my-claude installation docs, memory optimization",
|
||||
"author": {
|
||||
"name": "Maksim Kochetkov",
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "brewtools",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"description": "Brewtools - universal utilities for Claude Code: text optimization, humanization, secrets scanning",
|
||||
"author": {
|
||||
"name": "Maksim Kochetkov",
|
||||
|
||||
@@ -1,27 +1,16 @@
|
||||
---
|
||||
name: text-optimizer
|
||||
description: |
|
||||
Optimizes text/docs for LLM efficiency. Supports 4 modes: light, medium, standard (30-50% human-readable), deep (2-3x LLM-only with DICT+symbols). Triggers: "optimize prompt", "reduce tokens", "compress text", "too verbose", "deep compress", "standard compress".
|
||||
Optimizes text and docs for LLM token efficiency. Modes: light, medium, standard (30-50% human-readable), deep (2-3x LLM-only with DICT+symbols). Triggers: optimize prompt, reduce tokens, compress text, too verbose, deep compress, standard compress, token efficiency, shrink docs.
|
||||
|
||||
<example>
|
||||
Context: User wants to reduce token usage
|
||||
user: "Optimize this prompt for fewer tokens"
|
||||
assistant: "I'll analyze and compress it using text-optimize skill."
|
||||
<commentary>Explicit optimization request triggers this agent</commentary>
|
||||
<commentary>Explicit compression request - apply standard or deep mode.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
Context: Document is too verbose
|
||||
user: "This CLAUDE.md is too long"
|
||||
assistant: "I'll optimize it for LLM consumption."
|
||||
<commentary>Verbose document complaint triggers this agent</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
Context: User mentions token efficiency
|
||||
user: "Make this more token-efficient"
|
||||
assistant: "I'll apply token reduction techniques."
|
||||
<commentary>Token efficiency mention triggers this agent</commentary>
|
||||
<commentary>LLM-only file - deep mode with DICT + symbols.</commentary>
|
||||
</example>
|
||||
model: sonnet
|
||||
color: magenta
|
||||
|
||||
@@ -68,8 +68,8 @@ async function main() {
|
||||
const sessionShort = session_id?.slice(0, 8) || 'unknown';
|
||||
|
||||
let context = pluginRoot
|
||||
? `BT_PLUGIN_ROOT=${pluginRoot}\nbrewtools: active | session: ${sessionShort}`
|
||||
: `brewtools: active | session: ${sessionShort}`;
|
||||
? `BT_PLUGIN_ROOT=${pluginRoot}`
|
||||
: `brewtools: ${sessionShort}`;
|
||||
|
||||
context = await injectThinkShort(context, cwd, session_id);
|
||||
|
||||
|
||||
@@ -1,29 +1,6 @@
|
||||
---
|
||||
name: brewtools:agent-toggle
|
||||
description: |
|
||||
Disable or enable individual plugin agents by renaming <name>.md↔_<name>.md in the plugin cache, with persistent state (global or project scope). Use for hiding specific agents without uninstalling the plugin.
|
||||
|
||||
Triggers: "disable agent", "enable agent", "agent toggle", "hide agent", "agent-toggle", "toggle agent", "отключи агента", "выключи агента", "включи агента", "спрятать агента".
|
||||
|
||||
<example>
|
||||
user: "отключи агента brewtools:ssh-admin"
|
||||
<commentary>NL disable request — plugin=brewtools, name=ssh-admin, scope=global (default).</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "enable brewcode:developer for this project only"
|
||||
<commentary>Scope=project — write to .claude/brewtools/toggle-state.json.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "agent-toggle status"
|
||||
<commentary>Show merged table of disabled agents (project overrides global).</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "disable tester"
|
||||
<commentary>Ambiguous — no plugin prefix. Ask which plugin owns the agent before mutating.</commentary>
|
||||
</example>
|
||||
description: "Disables or enables individual plugin agents by renaming <name>.md and _<name>.md in the plugin cache. Persistent state: global (default) or project. Triggers: disable agent, enable agent, toggle agent, hide agent, agent-toggle."
|
||||
argument-hint: "[disable|enable|status|list|reapply|prune] <plugin>:<name> [--scope=global|project]"
|
||||
allowed-tools: Read, Bash, AskUserQuestion
|
||||
model: sonnet
|
||||
|
||||
@@ -1,29 +1,6 @@
|
||||
---
|
||||
name: brewtools:skill-toggle
|
||||
description: |
|
||||
Disable or enable individual plugin skills by renaming SKILL.md↔_SKILL.md in the plugin cache, with persistent state (global or project scope). Use for hiding specific skills without uninstalling the plugin.
|
||||
|
||||
Triggers: "disable skill", "enable skill", "skill toggle", "hide skill", "skill-toggle", "toggle skill", "отключи скил", "выключи скил", "включи скил", "спрятать скил".
|
||||
|
||||
<example>
|
||||
user: "отключи скил brewui:image-gen"
|
||||
<commentary>NL disable request — plugin=brewui, name=image-gen, scope=global (default).</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "enable brewcode:convention for this project only"
|
||||
<commentary>Scope=project — write to .claude/brewtools/toggle-state.json.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "skill-toggle status"
|
||||
<commentary>Show merged table of disabled skills (project overrides global).</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "disable image-gen"
|
||||
<commentary>Ambiguous — no plugin prefix. Ask which plugin owns the skill before mutating.</commentary>
|
||||
</example>
|
||||
description: "Disables or enables individual plugin skills by renaming SKILL.md and _SKILL.md in the plugin cache. Persistent state: global (default) or project. Triggers: disable skill, enable skill, toggle skill, hide skill, skill-toggle."
|
||||
argument-hint: "[disable|enable|status|list|reapply|prune] <plugin>:<name> [--scope=global|project]"
|
||||
allowed-tools: Read, Bash, AskUserQuestion
|
||||
model: sonnet
|
||||
|
||||
@@ -1,44 +1,6 @@
|
||||
---
|
||||
name: brewtools:think-short
|
||||
description: |
|
||||
Toggle terse-output mode for Claude Code — injects brief directives into main conversation (SessionStart) and sub-agent prompts (PreToolUse:Task) to reduce preamble, filler, and AI artifacts. Profiles: light / medium / aggressive. State is global or project-scoped with merge precedence (project > global).
|
||||
|
||||
Triggers: "think-short", "терсный режим", "короче отвечай", "reduce tokens mode", "be terse", "think shorter", "токен-экономия", "think-short on", "think-short off", "think-short profile", "think-short status", "think-short blacklist", "включи терсный", "выключи терсный", "сделай покороче", "меньше воды", "уровень 1", "уровень 2", "уровень 3", "level 1", "level 2", "level 3", "макс", "максимум", "максимально", "агрессивный режим", "средний режим", "лёгкий режим", "легкий режим".
|
||||
|
||||
<example>
|
||||
user: "/brewtools:think-short on"
|
||||
<commentary>Structural `on` — default scope=project, AskUserQuestion to confirm Project/Global.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "включись максимально"
|
||||
<commentary>NL combo — resolves to `on` + `profile aggressive`. Log the resolution, execute both.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "think-short status"
|
||||
<commentary>Print effective state, source, state-file existence, plugin defaults, env override, last 10 log lines.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "покороче"
|
||||
<commentary>Ambiguous (matches `on` OR `profile light`?) — AskUserQuestion to disambiguate.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "уровень 3"
|
||||
<commentary>Bare level → resolves to `profile aggressive`. Default scope=project (silent). Log resolution.</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "level 2"
|
||||
<commentary>Bare English level → resolves to `profile medium`. Default scope=project (silent).</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "макс"
|
||||
<commentary>Shorthand for aggressive → resolves to `profile aggressive`. Default scope=project (silent).</commentary>
|
||||
</example>
|
||||
description: "Toggles terse-output mode for Claude Code - injects brief directives into main conversation (SessionStart) and sub-agent prompts (PreToolUse:Task) to cut preamble and filler. Profiles: light, medium, aggressive. Scope: global or project (project overrides). Triggers: think-short, be terse, think shorter, reduce tokens mode, level 1/2/3."
|
||||
argument-hint: "[on|off|profile <light|medium|aggressive>|status|blacklist add|remove <agent>] [--scope global|project]"
|
||||
allowed-tools: Read, Bash, AskUserQuestion
|
||||
model: sonnet
|
||||
|
||||
@@ -1,20 +1,15 @@
|
||||
Be terse. ASCII only — no em-dash, no smart quotes.
|
||||
No preamble ("Let me...", "I'll...", "Sure!").
|
||||
No closing fluff ("Hope this helps!", "Let me know!").
|
||||
No sycophantic affirmations ("Great question!", "Absolutely!").
|
||||
No restatement of the question before answering.
|
||||
No unsolicited alternatives / refactoring suggestions.
|
||||
No "as an AI" framing, no disclaimers.
|
||||
Be terse. ASCII only - no em-dash, no smart quotes.
|
||||
No preamble ("Let me...", "Sure!").
|
||||
No closing fluff, sycophancy, disclaimers, or "as an AI".
|
||||
No restatement of question. No unsolicited alternatives.
|
||||
Results first. Reasoning only if explicitly asked.
|
||||
Minimize chain-of-thought verbalization in output.
|
||||
Prefer Edit over Write. Diff over full file.
|
||||
Test before declaring done.
|
||||
Prefer Edit over Write. Diff over full file. Test before declaring done.
|
||||
User instructions always override these rules.
|
||||
|
||||
Tools: Grep/Glob before Read — never scan whole files to find a symbol.
|
||||
Bundle related edits per file into fewer calls. replace_all=true beats N identical Edits.
|
||||
Parallel independent tool calls (reads, greps, different files) in one message.
|
||||
Don't re-Read a file you just edited — Edit returns the post-edit snippet.
|
||||
Grep/Glob before Read.
|
||||
Bundle edits per file. replace_all=true beats repeated Edits.
|
||||
Parallel calls (reads, greps, different files) in one message.
|
||||
Don't re-Read a file you just edited.
|
||||
|
||||
Think before acting: reason through the whole edit set first, then execute — no see/edit/see/edit loops.
|
||||
Plan the full edit set, then execute.
|
||||
If a task touches N call-sites, gather all N via Grep first, then issue parallel Edits.
|
||||
|
||||
@@ -1,3 +1,3 @@
|
||||
Be terse. No preamble, no filler, no AI phrasings.
|
||||
Results first, reasoning only if asked.
|
||||
Think through the whole edit set first, then execute — no see/edit/see/edit loops.
|
||||
Plan the full edit set, then execute.
|
||||
|
||||
@@ -1,9 +1,6 @@
|
||||
Be terse. No preamble, no filler.
|
||||
Skip "Let me...", "I'll now...", "Great question".
|
||||
Direct answers — results first, reasoning only if asked.
|
||||
Minimize chain-of-thought verbalization.
|
||||
No sycophancy, no unsolicited alternatives.
|
||||
Be terse. No preamble, filler, sycophancy, or unsolicited alternatives.
|
||||
Results first, reasoning only if asked.
|
||||
|
||||
Tools: Grep/Glob before Read. Edit over Write. Parallel independent tool calls in one message.
|
||||
Grep/Glob before Read. Edit over Write. Parallel calls in one message.
|
||||
Don't re-Read a file you just edited.
|
||||
Think through the edit set first, then execute — no see/edit/see/edit loops.
|
||||
Plan the full edit set, then execute.
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "brewui",
|
||||
"version": "3.7.6",
|
||||
"version": "3.7.7",
|
||||
"description": "Brewui - UI/visual/creative tools for Claude Code: AI image generation, design-to-code conversion",
|
||||
"author": {
|
||||
"name": "Maksim Kochetkov",
|
||||
|
||||
@@ -1,17 +1,6 @@
|
||||
---
|
||||
name: glm-zai-specialist
|
||||
description: |
|
||||
Z.ai GLM API expert — vision requests, model selection, rate limits, response parsing for design-to-code pipeline. Triggers: "zai api", "glm request", "z.ai", "send to glm", "glm vision", "glm model", "design to code api", "glm-5v", "glm-4.6v".
|
||||
|
||||
<example>
|
||||
user: "Send this screenshot to GLM for design-to-code conversion"
|
||||
<commentary>Direct GLM API request with vision input — core specialist task</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "GLM is returning 429 errors, fix the request"
|
||||
<commentary>API troubleshooting with rate limit handling — specialist domain</commentary>
|
||||
</example>
|
||||
description: "Z.ai GLM API expert - vision requests, model selection, rate limits, response parsing for design-to-code pipeline. Triggers: zai api, glm request, z.ai, send to glm, glm vision, glm-5v, glm-4.6v."
|
||||
model: sonnet
|
||||
color: cyan
|
||||
tools: Read, Write, Edit, Bash, Glob, Grep
|
||||
|
||||
@@ -14,8 +14,8 @@ async function main() {
|
||||
const sessionShort = session_id?.slice(0, 8) || 'unknown';
|
||||
|
||||
const context = pluginRoot
|
||||
? `BU_PLUGIN_ROOT=${pluginRoot}\nbrewui: active | session: ${sessionShort}`
|
||||
: `brewui: active | session: ${sessionShort}`;
|
||||
? `BU_PLUGIN_ROOT=${pluginRoot}`
|
||||
: `brewui: ${sessionShort}`;
|
||||
|
||||
output({
|
||||
systemMessage: `brewui: ${pluginRoot} | session: ${sessionShort}`,
|
||||
|
||||
@@ -1,23 +1,6 @@
|
||||
---
|
||||
name: brewui:glm-design-to-code
|
||||
description: |
|
||||
Use for generates frontend code from design screenshots, mockups, text descriptions, HTML files, or URLs using external GLM vision API (not Claude). Three modes: CREATE, REVIEW, FIX. Supports HTML, React, Flutter output.
|
||||
Triggers: "convert screenshot to code", "design to code", "mockup to code", "generate frontend from image", "turn design into React", "screenshot to HTML", "d2c"
|
||||
|
||||
<example>
|
||||
user: "Convert this screenshot to React components"
|
||||
<commentary>CREATE mode — user has a design image and wants framework code</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "Turn my mockup into a working landing page"
|
||||
<commentary>CREATE mode — natural request for design-to-code generation</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "Review how well the generated code matches the original design"
|
||||
<commentary>REVIEW mode — user wants visual comparison of design vs result</commentary>
|
||||
</example>
|
||||
description: "Generates frontend code from design screenshots, mockups, text descriptions, HTML files, or URLs using external GLM vision API. Modes: CREATE, REVIEW, FIX. Output: HTML, React, Flutter. Triggers: design to code, screenshot to code, mockup to code, d2c, generate frontend from image."
|
||||
user-invocable: true
|
||||
argument-hint: "[input] [--framework html|react|flutter|custom] [--profile max|optimal|efficient] [--provider zai|openrouter] [--model MODEL_ID] [--output dir] [--review original.png result.png] [--fix 'feedback'] [--fix --review-file review.json]"
|
||||
allowed-tools: [Read, Write, Edit, Glob, Grep, Bash, AskUserQuestion]
|
||||
|
||||
@@ -1,24 +1,6 @@
|
||||
---
|
||||
name: brewui:image-gen
|
||||
description: |
|
||||
Generates AI images via multiple providers (OpenRouter Gemini/GPT-5, Z.ai GLM-image, Google Imagen 4, OpenAI DALL-E 3) with anti-AI-slop controls.
|
||||
Modes: generate (default), edit, config, update.
|
||||
Triggers: "generate image", "create image", "make image", "AI image", "image-gen", "og image", "blog image", "illustration"
|
||||
|
||||
<example>
|
||||
user: "Generate an OG image for my blog post about dark mode"
|
||||
<commentary>generate mode — text prompt to image</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "Edit this image to add warm lighting"
|
||||
<commentary>edit mode — modify existing image</commentary>
|
||||
</example>
|
||||
|
||||
<example>
|
||||
user: "/image-gen --config"
|
||||
<commentary>config mode — set up API keys</commentary>
|
||||
</example>
|
||||
description: "Generates AI images via OpenRouter (Gemini/GPT-5), Z.ai GLM-image, Google Imagen 4, or OpenAI DALL-E 3 with anti-AI-slop controls. Modes: generate, edit, config, update. Triggers: generate image, create image, AI image, og image, blog image, illustration."
|
||||
user-invocable: true
|
||||
argument-hint: "[prompt] [--edit image.png 'instructions'] [--config] [--update] [--service gemini|openrouter|openai] [--style photo|illustration|art] [--count N] [--output dir] [--size WxH]"
|
||||
allowed-tools: [Read, Write, Edit, Bash, Glob, Grep, AskUserQuestion, WebSearch]
|
||||
|
||||
Reference in New Issue
Block a user