How do I install the Btr Capture skill?

Run npx skillmds add majiayu000/btr-capture in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Btr Capture safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: reads secrets. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Btr Capture?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Btr Capture free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Btr Capture?

majiayu000 (@majiayu000) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Btr Capture by majiayu000 — Agent Skill

Name: Btr Capture
Rating: 4.2 (5 reviews)
Author: majiayu000

Btr Capture

Save to LOCAL BTR context tree (NOT ByteRover/brv). Capture code snippets, patterns, and context from the current conversation and store them in the BTR context tree. PROACTIVE TRIGGER: Automatically suggest saving when the user confirms something works ("That works!", "Perfect!", "It's working now", "Thanks, that fixed it!"), after fixing a bug, making an architecture decision, establishing a pattern, or solving a complex problem. Also use when the user explicitly says "save this", "remember this pattern", "capture this context", "save to BTR", "save this to BTR", or "add to BTR".

majiayu000 de4c65a 2 files · 4.9 KB Updated Sep 9, 2026 567 repo stars

File contents

BTR Capture

⚠️ CRITICAL: BTR ≠ ByteRover

This skill uses btr (local context tree), NOT brv (ByteRover CLI).

Command	Tool	Syntax
✓ CORRECT	`btr`	`btr curate <domain> <topic> --content "..."`
✗ WRONG	`brv`	Different tool, different syntax, requires auth

PREFER MCP tools when available:

mcp__btr__curate_context - Structured, type-safe
mcp__btr__query_context - Validated search

Only use Bash btr commands if MCP tools are unavailable.

Capture and store valuable context for future retrieval.

PROACTIVE BEHAVIOR (CRITICAL)

DO NOT wait for the user to say "save this" - you should proactively suggest saving valuable context.

Trigger Conditions

After ANY of these events, ASK the user if they want to save the context:

User confirms code works - "That works!", "Perfect!", "It's working now"
Bug was fixed - Successfully resolved an issue after debugging
Architecture decision made - Discussed and decided on a design approach
New pattern established - Created a reusable pattern, utility, or component
Complex problem solved - Figured out a non-obvious solution
Configuration established - Set up tooling, environment, or integration

Example Flow

User: "That fixed the authentication issue, thanks!"

Claude: "Great! The JWT refresh token rotation pattern we implemented
could be valuable for future auth work. Save to BTR?

I'd capture:
- Domain: auth
- Topic: jwt-refresh-rotation
- Key details: The rotation logic, error handling, and token invalidation"

Proactive Check-In

Every 10-15 messages during active development, consider:

Have we established patterns worth preserving?
Did we make decisions that should be documented?
Is there context that would help future sessions?

Preferred Method

FIRST: Use MCP tools if available

mcp__btr__curate_context(domain="auth", topic="jwt-flow", content="...", tags=["security"])

FALLBACK: Use btr CLI via Bash

btr curate auth jwt-flow --content "..." --tags security

NEVER: Use brv (different product entirely)

Quick Start

btr curate <domain> <topic> --content "<content>" [--tags tag1,tag2]

Instructions

Identify the content to capture (code, explanation, pattern)
Determine appropriate domain (e.g., auth, api, database, frontend, testing)
Generate a descriptive topic name (kebab-case, e.g., jwt-validation, error-handling)
Extract or ask for relevant tags
Run the CLI command to save the context
Confirm successful capture to user

Domain Suggestions

auth - Authentication, authorization, sessions
api - REST endpoints, GraphQL, rate limiting
database - Queries, migrations, connection pooling
frontend - Components, state management, styling
testing - Test patterns, mocking, fixtures
devops - Deployment, CI/CD, monitoring
architecture - Design decisions, patterns

Examples

Save a code pattern

User: "Save this JWT validation pattern"

btr curate auth jwt-validation --content "..." --tags security,tokens,middleware

Save a design decision

User: "Remember why we chose Redis for caching"

btr curate architecture caching-decision --content "..." --tags redis,decisions,performance

For more examples, see examples.md.

---name: btr-capturedescription: Save to LOCAL BTR context tree (NOT ByteRover/brv). Capture code snippets, patterns, and context from the current conversation and store them in the BTR context tree. PROACTIVE TRIGGER: Automatically suggest saving when the user confirms something works ("That works!", "Perfect!", "It's working now", "Thanks, that fixed it!"), after fixing a bug, making an architecture decision, establishing a pattern, or solving a complex problem. Also use when the user explicitly says "save this", "remember this pattern", "capture this context", "save to BTR", "save this to BTR", or "add to BTR".--- # BTR Capture ## ⚠️ CRITICAL: BTR ≠ ByteRover **This skill uses `btr` (local context tree), NOT `brv` (ByteRover CLI).** | Command | Tool | Syntax | |---------|------|--------| | ✓ CORRECT | `btr` | `btr curate <domain> <topic> --content "..."` | | ✗ WRONG | `brv` | Different tool, different syntax, requires auth | **PREFER MCP tools when available:** - `mcp__btr__curate_context` - Structured, type-safe - `mcp__btr__query_context` - Validated search Only use Bash `btr` commands if MCP tools are unavailable. Capture and store valuable context for future retrieval. ## PROACTIVE BEHAVIOR (CRITICAL) **DO NOT wait for the user to say "save this"** - you should proactively suggest saving valuable context. ### Trigger Conditions After ANY of these events, ASK the user if they want to save the context: 1. **User confirms code works** - "That works!", "Perfect!", "It's working now" 2. **Bug was fixed** - Successfully resolved an issue after debugging 3. **Architecture decision made** - Discussed and decided on a design approach 4. **New pattern established** - Created a reusable pattern, utility, or component 5. **Complex problem solved** - Figured out a non-obvious solution 6. **Configuration established** - Set up tooling, environment, or integration ### Example Flow ``` User: "That fixed the authentication issue, thanks!" Claude: "Great! The JWT refresh token rotation pattern we implemented could be valuable for future auth work. Save to BTR? I'd capture: - Domain: auth - Topic: jwt-refresh-rotation - Key details: The rotation logic, error handling, and token invalidation" ``` ### Proactive Check-In Every 10-15 messages during active development, consider: - Have we established patterns worth preserving? - Did we make decisions that should be documented? - Is there context that would help future sessions? ## Preferred Method 1. **FIRST**: Use MCP tools if available ``` mcp__btr__curate_context(domain="auth", topic="jwt-flow", content="...", tags=["security"]) ``` 2. **FALLBACK**: Use `btr` CLI via Bash ```bash btr curate auth jwt-flow --content "..." --tags security ``` 3. **NEVER**: Use `brv` (different product entirely) ## Quick Start ```bash btr curate <domain> <topic> --content "<content>" [--tags tag1,tag2] ``` ## Instructions 1. Identify the content to capture (code, explanation, pattern) 2. Determine appropriate domain (e.g., auth, api, database, frontend, testing) 3. Generate a descriptive topic name (kebab-case, e.g., jwt-validation, error-handling) 4. Extract or ask for relevant tags 5. Run the CLI command to save the context 6. Confirm successful capture to user ## Domain Suggestions - `auth` - Authentication, authorization, sessions - `api` - REST endpoints, GraphQL, rate limiting - `database` - Queries, migrations, connection pooling - `frontend` - Components, state management, styling - `testing` - Test patterns, mocking, fixtures - `devops` - Deployment, CI/CD, monitoring - `architecture` - Design decisions, patterns ## Examples ### Save a code pattern User: "Save this JWT validation pattern" ```bash btr curate auth jwt-validation --content "..." --tags security,tokens,middleware ``` ### Save a design decision User: "Remember why we chose Redis for caching" ```bash btr curate architecture caching-decision --content "..." --tags redis,decisions,performance ``` For more examples, see [examples.md](examples.md).

Frequently asked questions