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.
npx agentmods add skills/nweii/agent-stuff/semantic-compressionnpx skills add nweii/agent-stuff --skill semantic-compressiongit clone --depth 1 https://github.com/nweii/agent-stuffWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00056 | $0.01130 |
| Opus 5 | $0.00028 | $0.00565 |
| Sonnet 5 | $0.00011 | $0.00226 |
| Haiku 4.5 | $0.00006 | $0.00113 |
Grade A, and why
semantic-compression 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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Semantic compression
Find the smallest expression that carries the meaning the user needs. Compression may produce a tighter explanation, a vivid phrase, one concrete word, an established term, or a coined expression. Success means preserving the distinctions, implications, and texture that matter in context.
Process
1. Establish the semantic payload
Read the surrounding conversation, source material, and writing guidelines before asking questions. Identify:
- what the expression must literally mean
- what it should imply without stating outright
- its emotional texture and degree of force
- the relationships it must preserve, such as direction, causality, reciprocity, or recursion
- its audience, register, and role in the larger passage
- any distinctions that nearby words or concepts already carry
Infer only what the available context supports. If a missing detail could materially change the result, ask one specific question at a time until the required payload is clear. Do not ask the user to repeat context already available.
This step is complete when you can state what the compressed expression must preserve and what it may discard.
2. Separate meaning from wording
Restate the payload internally without borrowing the source construction. Identify accidental bulk: modifiers compensating for a weak noun or verb, repeated implications, abstractions hiding a concrete action, qualifications that no longer change the claim, or several adjacent words pointing at one available concept. Look for a word whose ordinary associations absorb the work of a modifier or explanation.
Keep genuine qualifications. A shorter expression that broadens, hardens, sanitizes, or prettifies the claim is a mistranslation.
This step is complete when every removable part has been distinguished from meaning-bearing detail.
3. Generate candidates at useful scales
Search beyond synonyms. Depending on the material, try:
- a direct, tighter explanation
- a concrete verb or noun that absorbs its modifiers
- a vivid phrase that recruits familiar associations
- an established term that names the full concept
- a coined expression when existing language misses an important combination
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.
- yesterday First seen · 109 lines · 56 tokens per session scan A ca0182f2e50b
semantic-compression is a skill published in the GitHub repository nweii/agent-stuff (8 stars, last pushed 13d ago), licensed MIT. It adds 56 tokens to every session and 1,130 once invoked, about $0.0003 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.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…