agentleFS
Sign inStart free
GauravRatnawatGauravRatnawat/claude-cost-optimizer/templates/CLAUDE.md

claude-cost-optimizer / templates

CLAUDE.md0 starsChanged 7 months agoMITView on GitHub

You are configured for maximum cost efficiency while maintaining high quality output. Automatically save every conversation using the conversation-recorder skill after completing significant work or when the conversation reaches a natural stopping point. Do this proactively without being asked. This eliminates the need to re-explain context in new sessions, saving tokens. Use the /browse skill from gstack for all web browsing tasks. NEVER use other browser tools. Available gstack skills: - /plan-ceo-review - CEO-level project planning - /plan-eng-review - Engineering…

We split this CLAUDE.md into the rules it gives an agent

5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the /browse skill from gstack for all web browsing tasks.

    line 21
  2. Distinctive:

    If code-review-graph is installed, use these skills for efficient code reviews:

    line 39
  3. Distinctive:

    Stay focused: Avoid tangential explorations unless explicitly requested

    line 51
  4. Distinctive:

    Trust tools: Use specialized tools (Grep, Glob, Read) instead of bash commands

    line 52
  5. Distinctive:

    When working in Conductor workspaces, maintain focus on the workspace's specific project.

    line 60

≈ “Common” is an estimate: at least 1 in 5 of the 25,680 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

Global Instructions

Cost Optimization Strategy

You are configured for maximum cost efficiency while maintaining high quality output.

Model Usage

  • Complex planning tasks: Use full capabilities
  • Routine execution: Be concise and efficient
  • Current model: opusplan (automatically optimizes model selection)

Conversation Recording

Automatically save every conversation using the conversation-recorder skill after completing significant work or when the conversation reaches a natural stopping point. Do this proactively without being asked.

This eliminates the need to re-explain context in new sessions, saving tokens.

Web Browsing

Use the /browse skill from gstack for all web browsing tasks. NEVER use other browser tools.

Available gstack skills:

  • /plan-ceo-review - CEO-level project planning
  • /plan-eng-review - Engineering planning review
  • /review - Pre-landing PR review (catches SQL safety, LLM trust boundaries, etc.)
  • /qa - Systematic QA testing (diff-aware mode on feature branches)
  • /ship - Complete ship workflow (merge, test, changelog, commit, push, PR)
  • /browse - Headless browser testing (~100ms per command)
  • /retro - Weekly engineering retrospective
  • /setup-browser-cookies - Configure browser authentication

Code Navigation

GitNexus pre-hooks are configured for intelligent code search. Trust the hook suggestions for efficient navigation.

Code Review with Knowledge Graphs

If code-review-graph is installed, use these skills for efficient code reviews:

  • /review-delta - Review only changes since last commit (token-efficient)
  • /review-pr - Review PR or branch diff with structural context
  • /build-graph - Build/update the code knowledge graph

These tools provide blast-radius analysis and impact detection for safer code changes.

Efficiency Guidelines

  1. Reuse context: Reference previous conversations when available
  2. Batch operations: Combine related queries when possible
  3. Use skills: Leverage /review, /qa, /ship instead of manual workflows
  4. Stay focused: Avoid tangential explorations unless explicitly requested
  5. Trust tools: Use specialized tools (Grep, Glob, Read) instead of bash commands

Permission Strategy

Common safe commands are pre-authorized. For new commands, explain the purpose clearly when requesting permission.

Workspace Awareness

When working in Conductor workspaces, maintain focus on the workspace's specific project. Avoid pulling in context from unrelated work.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free