Does it overwrite my manual changes when updating?

No, the update process is designed to merge changes, preserving human-curated descriptions while flagging potentially stale entries for review.

How does it keep the token count low?

The skill follows strict inclusion/exclusion criteria, omitting generated files, boilerplate, and information that an agent can easily infer from filenames alone.

Which programming languages does AgentMap support?

It includes built-in heuristics for Python, JavaScript/TypeScript, Rust, Go, Java, C#, Ruby, Elixir, C/C++, and generic Make-based projects.

Is this a replacement for standard project documentation?

No, it is a navigation aid meant to point an agent to the right files to read, not a substitute for explaining how the code functions in detail.

What is an .agentmap.yaml file?

It is a lean YAML routing table that tells AI agents where important project components are located without providing excessive, token-heavy details.

AgentMap Codebase Mapper

Name: AgentMap Codebase Mapper
Author: jinglemansweep

byjinglemansweep

0•

学習とドキュメンテーション

Generates and maintains a lean .agentmap.yaml orientation map to help AI agents navigate complex codebases efficiently.

AgentMap is a specialized skill designed to create a high-efficiency "routing table" for AI agents within a repository. By generating a structured .agentmap.yaml file, it provides Claude and other agents with immediate high-level context—including entry points, core abstractions, and project conventions—without wasting valuable context window tokens on redundant file listings or deep documentation. It automates the reconnaissance of project structures across various ecosystems like Python, Node.js, Go, and Rust, ensuring that any AI agent can quickly identify where to start reading and how to execute essential project tasks.

主な機能

01Smart update logic that preserves human-curated annotations while refreshing structural data

02Supports monorepo architectures through nested sub-maps and pointers

03Automated codebase reconnaissance across multiple programming ecosystems and build tools

040 GitHub stars

05Generates token-efficient YAML maps focused on navigation rather than exhaustive documentation

06Identifies key architectural symbols, entry points, and project-specific conventions

ユースケース

01Standardizing project orientation maps for teams utilizing agentic development workflows

02Onboarding an AI agent to a large, unfamiliar repository to reduce navigation errors and hallucinations

03Maintaining lean architectural summaries that fit within tight LLM context windows for better performance

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jinglemansweep/agent-resources agentmap-generate

For use in Claude.ai and ChatGPT

主な機能

01Smart update logic that preserves human-curated annotations while refreshing structural data

02Supports monorepo architectures through nested sub-maps and pointers

03Automated codebase reconnaissance across multiple programming ecosystems and build tools

040 GitHub stars

05Generates token-efficient YAML maps focused on navigation rather than exhaustive documentation

06Identifies key architectural symbols, entry points, and project-specific conventions

ユースケース

01Standardizing project orientation maps for teams utilizing agentic development workflows

02Onboarding an AI agent to a large, unfamiliar repository to reduce navigation errors and hallucinations

03Maintaining lean architectural summaries that fit within tight LLM context windows for better performance