map

A codebase-mapping tool that sends research agents through several waves to create focused context documents. A codebase is the full set of source files and project materials, and context documents summarise how it is organised and works.

In plain words
What is it for?
Use it to map a repository’s technology stack, structure, architecture, conventions, and other project patterns into reference documents.
Why use it?
It reduces duplicated investigation by having later research build on earlier findings. The resulting documents make an unfamiliar project easier to navigate.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/qgolem/orc/map
Any agent
npx skills add qGolem/orc --skill map
Clone the repo
git clone --depth 1 https://github.com/qGolem/orc

Made for: Claude Code, Codex.

Per session 16 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 7,350 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00016 $0.07350
Opus 5 $0.00008 $0.03675
Sonnet 5 $0.00003 $0.01470
Haiku 4.5 $0.00002 $0.00735

Measured 2d ago against content hash 53445b892edb, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

map scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/map/SKILL.md · 918 lines

How it starts

The opening of the file, as written. The whole thing — 918 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Each wave builds on prior waves — later agents read earlier documents instead of re-exploring the same ground. This eliminates duplication and produces focused, non-overlapping output.

Output: .claude/context/ folder with 16 structured documents, each ≤200 lines.

Why waves: Parallel-only deployment causes agents to independently discover the same structure, stack, and patterns — producing 3-4 files that overlap on architecture, naming, validation, etc. Waves let foundation docs (STACK, STRUCTURE) be written first, then referenced by all subsequent agents.

Why ≤200 lines: Documents are reference material loaded into context. Bloated files waste tokens. Prefer tables over prose, examples over explanations. If a topic is covered in another doc, write "See [DOC].md" instead of repeating.

Upstream enrichment: Wave 2+ agents may update prior-wave documents if they discover information that would improve output for later waves. For example, a Wave 2 architecture agent might add entry points it discovered to STRUCTURE.md, or a Wave 3 patterns agent might add a convention it found to CONVENTIONS.md. Append to the relevant section — never delete or rewrite existing content.

Why dedicated mapper agents:

  • Fresh context per domain (no token contamination)
  • Agents write documents directly (no context transfer back to orchestrator)
  • Orchestrator only summarizes what was created (minimal context usage)

Always include file paths: Documents are reference material for Claude when planning/executing. Always include actual file paths formatted with backticks: src/services/user.ts.

Target directory: .claude/context/

This command can run:

  • Before /orc (brownfield codebases) -- creates codebase map first
  • After /orc (greenfield codebases) -- updates codebase map as code evolves
  • Anytime to refresh codebase understanding

<when_to_use> Use orc-map for:

  • Brownfield projects before planning (understand existing code first)
  • Refreshing codebase map after significant changes
  • Onboarding to an unfamiliar codebase
  • Before major refactoring (understand current state)

Skip orc-map for:

  • Greenfield projects with no code yet (nothing to map)
  • Trivial codebases (<5 files) </when_to_use>

Step 1: Check existing maps

ls -la .claude/context/ 2>/dev/null && echo "has_maps=true" || echo "has_maps=false"

If context/ already exists:

Present options:

AskUserQuestion:
  header: "Existing"
  question: ".claude/context/ already exists. What should we do?"
  options:
    - label: "Refresh"
      description: "Delete existing and remap codebase"
    - label: "Skip"
      description: "Use existing codebase map as-is"

If "Refresh": Delete context/, continue to step 2 If "Skip": Exit workflow

If doesn't exist: Continue to step 2.


Step 2: Create directory structure

mkdir -p .claude/context

Expected output files (16 documents across 4 waves):

Wave 1 — Foundation:

  • STACK.md, STRUCTURE.md, HISTORY.md

Wave 2 — Architecture & Conventions:

  • ARCHITECTURE.md, CONVENTIONS.md, INTEGRATIONS.md, REPO.md

Wave 3 — Deep Analysis:

  • PATTERNS.md, TYPES.md, TESTING.md, CODEMAP.md, UX-PATTERNS.md

Wave 4 — Cross-Cutting:

  • BOUNDARIES.md, FLOWS.md, CONCERNS.md, TEST-QUALITY.md

Step 3a: Wave 1 — Foundation (3 parallel agents)

These are discovery agents with no prior context. They establish the facts that all later waves reference.

If $ARGUMENTS contains a prompt, append it to each agent's prompt as a <focus_hint> block.

Agent 1: Stack

Task(
  subagent_type="orc:orc-codebase-mapper",
  run_in_background=true,
  description="Map codebase tech stack",
  prompt="Focus: tech stack

Analyze this codebase for its technology stack and configuration.

Write to .claude/context/:
- STACK.md (max 150 lines)

Include:
- Languages, runtime versions, package manager
- Frameworks and key dependencies (with versions)
- Build tools, bundlers, task runners
- Environment configuration model (env files, config patterns)
- Dev vs production differences

Use tables for dependency listings. Be concise.

<focus_hint>$ARGUMENTS</focus_hint>

Explore thoroughly. Write document directly. Return confirmation only."
)

Read the full file on GitHub · 918 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 918 lines · 0 tokens per session scan A 53445b892edb

Subscribe to this mod's changes

map is a skill published in the GitHub repository qGolem/orc (5 stars, last pushed 5mo ago), licensed MIT. It adds 16 tokens to every session and 7,350 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

foundry-config-setup

Resolve missing setup caused by a hardcoded Foundry project endpoint or model in a sample. Use when a sample fails because it uses a placeholder/hardcoded projectendpoint (for example "https://your-project.services.ai.azure.com") or a hardcoded model instead of reading them from the environment.

microsoft/agent-framework · 65 tokens

trigger-authoring-tasks

Covers writing backend Trigger.dev tasks with @trigger.dev/sdk: defining task() and schemaTask(), the run function and its ctx, retries, waits, queues and concurrency, idempotency keys, run metadata, logging, triggering other tasks (and the Result shape), scheduled/cron tasks, and the essentials of trigger.config.ts.…

triggerdotdev/trigger.dev · 115 tokens

background

Use when the user wants to see, inspect, cancel, or prune background agents fired during prior chain runs. Read/manage .hyperflow/background/registry.json and the per-agent output buffers at .hyperflow/background/ .md. Standalone — never auto-invoked. Trigger with /hyperflow:background, "list background agents"…

jeremylongshore/tons-of-skills-marketplace · 89 tokens

local-frontend-check

Smoke-test or verify UI behaviour on the local Jarvis Registry frontend running at http://localhost/gateway. Use for manual regression checks, bug-fix verification, and end-to-end confirmation of specific flows without running the automated test suite.

ascending-llc/jarvis-registry · 52 tokens

atmos-auth

Authentication and identity management: providers (SSO/SAML/OIDC/GCP/Atmos Pro), identities, keyring, identity chaining, login/exec/shell/console, and github/sts for private GitHub access.

cloudposse/atmos · 48 tokens

brooks-sweep

Full-sweep mode: runs a unified analysis across all quality dimensions — code decay, architecture, tech debt, and test quality — then applies fixes directly to the codebase. Safe changes are auto-applied; risky changes are confirmed before execution. Drawing on twelve classic engineering books. Triggers when: user…

hyhmrright/brooks-lint · 178 tokens