Construction Claude Cowork Project Indexer — AI skill for Claude Code
Co-work skill that builds a full AI context system for construction projects.
How to install Construction Claude Cowork Project Indexer
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open hamzaabduljabbar/construction-claude-cowork-project-indexer and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Construction Claude Cowork Project Indexer does
Co-work skill that builds a full AI context system for construction projects. 20-40x fewer tokens. Consistent answers across every chat.
Alternatives in AI
README
project-indexer
A Co-work skill that runs once on any new construction project and builds the full 3-layer context system. After it runs, every chat in that project uses 20–40x fewer tokens and gets consistent, accurate answers.
Built for construction firms using Claude Co-work. Pairs with the [setup guide](docs/co-work-setup-guide.pdf) which covers the full 3-layer system, business context setup, prompting rules, and model selection.
What it does
Points Claude at a project folder. Claude reads every document — drawings, contract, specification, programme, BOQ — and builds four lightweight index files that every future chat reads instead of the raw PDFs.
YourProject/
└── 0.ai-context/ ← created by this skill
├── claude.md ← rules and architecture, read every chat
├── project.md ← full project snapshot: contract, parties, dates
├── drawings.md ← master drawing register and symbol library
├── memory.md ← running project log, updated each chat
└── processed/
├── SHEET-001.txt ← text layer extracted from each drawing sheet
├── SHEET-001.png ← rendered image of each sheet
└── ...
A raw construction drawing set costs 60,000–70,000 tokens every time Claude reads it. After this runs, the same drawing question costs 5,000–8,000 tokens because Claude reads `drawings.md` and one specific sheet file rather than the full PDF.
Repo structure
project-indexer/
├── SKILL.md # The full skill — 9-phase process + query protocol
├── scripts/
│ └── process_drawings.py # Splits PDFs, extracts text layer, renders PNGs
├── templates/
│ ├── claude-md-template.md # Template for the claude.md context file
│ └── project-md-template.md # Template for the project.md snapshot file
├── references/
│ └── document-types.md # What to extract from each construction doc type
└── docs
Related Skills
Lm Resizer
Rust-native context compression for Claude Code, Codex & MCP agents: filters & compresses noisy tool output (t
AI Assistant Instructions
A comprehensive, vendor-agnostic framework for consistent AI-assisted development workflows - standardized ins
Tsubasa
Your engineering captain: a repo-resident AI persona with a knowledge graph stored in git. Agents cite inciden
Scrooge Mode
Same answer, fewer tokens — KO-first LLM output-compression skill for Claude Code & Codex. A Korean-native cav
TokenMasterX
TokenMaster makes AI coding agents understand your codebase once and remember it - up to 7.8x fewer tokens.
Imperal MCP
An exoskeleton for your AI coding agent - it remembers everything, sees your whole codebase, can undo any chan
Related Agents
Glados
GLaDOS, Stef's factory chat at the work root, opened with factory glados. Starts standalone jobs with the fact
Part 19: Repowise — Codebase Intelligence For Coding Agents
60% fewer tokens. 4x faster. Your coding agents stop guessing and start knowing. ---
Lean Reader
Read-only reader, collector, or verifier stage that runs without CLAUDE.md and the memory index (18.8k fewer t