caveman-compress

Name: caveman-compress
Author: juliusbrussee

byjuliusbrussee· 57 skills

Installs

268k

Stars

93k

Forks

5.3k

How to install caveman-compress?

juliusbrussee/caveman/caveman-compress

$npx -y skills add juliusbrussee/caveman --skill caveman-compress

Installs into the current project.

›Prefer a prompt? Paste this to your agent

Use this skill

Run `npx skills use "https://github.com/juliusbrussee/caveman" --skill "juliusbrussee/caveman/caveman-compress"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.

Use the whole pack

Use the skills in "https://github.com/juliusbrussee/caveman" that are relevant to the current task. Run `npx skills add "https://github.com/juliusbrussee/caveman"` and select the relevant skills, then follow their instructions.

Files · 1

View on GitHub

SKILL.md

1	# Caveman Compress
2
3	## Purpose
4
5	Compress natural language files (CLAUDE.md, todos, preferences) into caveman-speak to reduce input tokens. Compressed version overwrites original. Human-readable backup saved as `<filename>.original.md`.
6
7	## Trigger
8
9	`/caveman-compress <filepath>` or when user asks to compress a memory file.
10
11	## Process
12
13	1. The compression scripts live in `scripts/` (adjacent to this SKILL.md). If the path is not immediately available, search for `scripts/__main__.py` next to this SKILL.md.
14
15	2. From the directory containing this SKILL.md, run:
16
17	python3 -m scripts <absolute_filepath>
18
19	3. The CLI will:
20	- detect file type (no tokens)
21	- call Claude to compress
22	- validate output (no tokens)
23	- if errors: cherry-pick fix with Claude (targeted fixes only, no recompression)
24	- retry up to 2 times
25	- if still failing after 2 retries: report error to user, leave original file untouched
26
27	4. Return result to user
28
29	## Compression Rules
30
31	### Remove
32	- Articles: a, an, the
33	- Filler: just, really, basically, actually, simply, essentially, generally
34	- Pleasantries: "sure", "certainly", "of course", "happy to", "I'd recommend"
35	- Hedging: "it might be worth", "you could consider", "it would be good to"
36	- Redundant phrasing: "in order to" → "to", "make sure to" → "ensure", "the reason is because" → "because"
37	- Connective fluff: "however", "furthermore", "additionally", "in addition"
38
39	### Preserve EXACTLY (never modify)
40	- Code blocks (fenced ``` and indented)
41	- Inline code (`backtick content`)
42	- URLs and links (full URLs, markdown links)
43	- File paths (`/src/components/...`, `./config.yaml`)
44	- Commands (`npm install`, `git commit`, `docker build`)
45	- Technical terms (library names, API names, protocols, algorithms)
46	- Proper nouns (project names, people, companies)
47	- Dates, version numbers, numeric values
48	- Environment variables (`$HOME`, `NODE_ENV`)
49
50	### Preserve Structure
51	- All markdown headings (keep exact heading text, compress body below)
52	- Bullet point hierarchy (keep nesting level)
53	- Numbered lists (keep numbering)
54	- Tables (compress cell text, keep structure)
55	- Frontmatter/YAML headers in markdown files
56
57	### Compress
58	- Use short synonyms: "big" not "extensive", "fix" not "implement a solution for", "use" not "utilize"
59	- Fragments OK: "Run tests before commit" not "You should always run tests before committing"
60	- Drop "you should", "make sure to", "remember to" — just state the action
61	- Merge redundant bullets that say the same thing differently
62	- Keep one example where multiple examples show the same pattern
63
64	CRITICAL RULE:
65	Anything inside ``` ... ``` must be copied EXACTLY.
66	Do not:
67	- remove comments
68	- remove spacing
69	- reorder lines
70	- shorten commands
71	- simplify anything
72
73	Inline code (`...`) must be preserved EXACTLY.
74	Do not modify anything inside backticks.
75
76	If file contains code blocks:
77	- Treat code blocks as read-only regions
78	- Only compress text outside them
79	- Do not merge sections around code
80
81	## Pattern
82
83	Original:
84	> You should always make sure to run the test suite before pushing any changes to the main branch. This is important because it helps catch bugs early and prevents broken builds from being deployed to production.
85
86	Compressed:
87	> Run tests before push to main. Catch bugs early, prevent broken prod deploys.
88
89	Original:
90	> The application uses a microservices architecture with the following components. The API gateway handles all incoming requests and routes them to the appropriate service. The authentication service is responsible for managing user sessions and JWT tokens.
91
92	Compressed:
93	> Microservices architecture. API gateway route all requests to services. Auth service manage user sessions + JWT tokens.
94
95	## Boundaries
96
97	- ONLY compress natural language files (.md, .txt, .typ, .typst, .tex, extensionless)
98	- NEVER modify: .py, .js, .ts, .json, .yaml, .yml, .toml, .env, .lock, .css, .html, .xml, .sql, .sh
99	- If file has mixed content (prose + code), compress ONLY the prose sections
100	- If unsure whether something is code or prose, leave it unchanged
101	- Original file is backed up as FILE.original.md before overwriting
102	- Never compress FILE.original.md (skip it)

6 picks

Type

Run `npx skills use "https://github.com/juliusbrussee/caveman" --skill "juliusbrussee/caveman/caveman-compress"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.

Use the skills in "https://github.com/juliusbrussee/caveman" that are relevant to the current task. Run `npx skills add "https://github.com/juliusbrussee/caveman"` and select the relevant skills, then follow their instructions.

caveman-compress

How to install caveman-compress?

Related

caveman-compress

How to install caveman-compress?

Related