Doc Condenser logo

Doc Condenser

Community
Mathews-Tom
doc-condenser

DEPRECATED: The base model handles document condensation and summarization natively at high quality. This skill no longer provides meaningful uplift. Retained for reference only.

Overview

PublisherMathews-Tom
Repositoryarmory
Skill namedoc-condenser
Stars
318
Forks
47
Bundled files
3
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by Mathews-Tom on GitHub. Read the source before you install it.

Installation

Install the Doc Condenser AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Mathews-Tom/armory.git /tmp/armory
mkdir -p .claude/skills
cp -r /tmp/armory/skills/doc-condenser .claude/skills/doc-condenser
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Doc Condenser in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Doc Condenser on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Doc Condenser is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

DEPRECATED — Modern Claude models condense and summarize technical documentation natively with comparable quality. The output format preferences encoded here (40% length cap, tables over prose, paths first) are too generic to justify skill overhead. Retained for archival reference only.

Document Condenser

Transform verbose technical documentation into concise, developer-focused specs.

Core Principles

  1. Paths first - Every file reference includes full/relative path
  2. Tables over prose - Use tables for metrics, coverage, file lists
  3. Code samples stay - Keep small, illustrative snippets; remove verbose examples
  4. Commentary, not explanation - Brief context sentences, not paragraphs
  5. One-line history - Reference legacy docs, don't preserve their content

Output Structure

markdown
# [path/to/output.md]

# [Title] - [Subtitle if needed]

## Purpose

[2-3 sentences: what this is, why it exists, key design principle]

## Status

[Table: metrics, rates, performance]

## Architecture Overview

[Optional diagram or brief flow description]
[Only if it aids understanding]

## Implementation Files

[Grouped by category with paths and one-line descriptions]

## [Domain-Specific Sections]

[Tables, code snippets, brief commentary as needed]

## Quick Reference

[Box or code block with key stats for scanning]

See assets/template.md for a copy-ready scaffold of this structure.

Transformation Rules

KEEP

  • File paths (always full or project-relative)
  • Metrics and measurements
  • Code snippets under 15 lines that illustrate patterns
  • Schema examples and data structures
  • Coverage/status tables

CONDENSE

  • Multi-paragraph explanations → 1-2 sentences
  • Verbose examples → representative snippet + "see X for more"
  • Implementation checklists → completion status table
  • Long rationales → single "Design principle: X" line
  • Code snippets longer than 20 lines → condense to the core pattern + a reference comment pointing to the source file
  • Rationale sections where the same point is restated across more than 3 sentences → collapse to one "Design principle:" line

REMOVE

  • Historical context beyond one reference line
  • Achieved/completed celebration language
  • Redundant explanations of the same concept
  • Step-by-step tutorials (link to them instead)
  • "What we learned" retrospectives

FORMAT

  • Use code blocks for paths and commands
  • Group related files under headers
  • Prefer tables over bullet lists for structured data
  • End with quick-reference block for scanning

Working with Existing Documents

When condensing an existing verbose doc:

  1. Identify the core purpose (first paragraph of output)
  2. Extract all file paths into grouped tables
  3. Preserve code samples that show patterns
  4. Convert prose sections to tables where possible
  5. Add single history reference line
  6. Verify no information loss on key technical details

Style Guide

See references/style-guide.md for detailed formatting rules, table patterns, and code sample guidelines.

Example Transformation

Before (verbose):

text
We have successfully achieved and EXCEEDED the original goals of this specification!
After many iterations and improvements, our automation rate reached 96.6% which is
above our target of 95%. The team worked hard on this and we're very proud...

After (concise):

text
**v31 PRODUCTION** | 96.6% automation (target: 95%)

Calibration Rules

  • Condensed output must not exceed 40% of source document length measured in words.
  • All file paths present in the original must appear in the condensed output — paths are never dropped.
  • Code blocks are never removed outright; reduce length by extracting the representative pattern and adding a source reference comment.

Error Handling

  • Source document has no clear purpose: ask for one sentence of context before condensing — do not infer a purpose and proceed.
  • A section contains items that are ambiguous between KEEP and REMOVE: default to KEEP and flag the section with a <!-- review: ambiguous --> comment in the output.
  • Condensed result loses required technical detail identified during verification: restore the omitted detail and re-measure against the 40% length cap; if the cap cannot be met, document the exception inline.
  • Source contains no file paths: skip the Implementation Files section entirely rather than generating placeholder paths.
  • Source is a non-text format (image, diagram, spreadsheet): report the format is unsupported and return without output.
  • Style guide conflicts with source formatting conventions: follow references/style-guide.md and note the override at the top of the output.

Limitations

  • Works on text documents only; images, diagrams, and binary files cannot be condensed.
  • Condensation ratio depends on source verbosity — highly structured sources yield less reduction.
  • Style guide deference (references/style-guide.md) takes precedence over source formatting, which can alter heading levels and table layouts.
  • Does not follow hyperlinks or fetch referenced external documents; referenced content is noted but not inlined.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Doc Condenser AI skill do?

DEPRECATED: The base model handles document condensation and summarization natively at high quality. This skill no longer provides meaningful uplift. Retained for reference only.

Why use Doc Condenser on TypingMind?

Because you install it once and use it with any model. Doc Condenser is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Doc Condenser in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Mathews-Tom/armory/tree/main/skills/doc-condenser. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Doc Condenser?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Doc Condenser?

As many as you like. As long as a model supports skills, you can use Doc Condenser with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Doc Condenser AI skill free?

Yes. It is published on GitHub by Mathews-Tom under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇