Menu
Same skill compression power in half the context — 975 tokens vs 2,500.
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
npx skills add https://github.com/live-neon/skills --skill skill-distiller-compressedThe command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
Based on SOC occupation classification
Showing SKILL.md
| name | Skill Distiller (Compressed) |
| version | 0.2.1 |
| description | Same skill compression power in half the context — 975 tokens vs 2,500. |
| author | Live Neon <lee@liveneon.ai> |
| homepage | https://github.com/live-neon/skills/tree/main/skill-distiller/compressed |
| repository | live-neon/skills |
| license | MIT |
| user-invocable | true |
| disable-model-invocation | true |
| emoji | 🗜️ |
| tags | ["compression","skills","optimization","context-window","token-reduction","lightweight","openclaw"] |
Self-compressed prose variant (~975 tokens, ~90% functionality, LLM-estimated). Full reference: ../SKILL.reference.md.
Role: Help users compress verbose skills to reduce context window usage Understands: Skills are verbose for human clarity but costly for context Approach: Identify section types, score importance, remove/shorten low-value sections Boundaries: Preserve functionality, report what was removed, never hide trade-offs Tone: Technical, precise, transparent about trade-offs
Data handling: All analysis uses your agent's configured model. No external APIs.
Activate when the user asks:
| Flag | Default | Description |
|---|---|---|
--mode | threshold | threshold (preserve X%), tokens (fit budget), oneliner |
--threshold | 0.9 | Functionality preservation target (0.0-1.0) |
--tokens | - | Target token count |
--provider | auto | ollama, gemini, openai (auto-detects) |
--verbose | false | Show section-by-section analysis |
--dry-run | false | Analyze without outputting |
Full options (--model, --debug-stages, --with-ci): see SKILL.reference.md
Threshold = semantic capability, not size ratio. A 0.9 threshold means 90% of agent behavior preserved, not 90% of lines kept. Judge by understanding, not metrics.
Parse into sections: Frontmatter, Headers, Code blocks, Lists, Prose.
| Type | Importance | Compressible? |
|---|---|---|
| TRIGGER | 1.0 | No |
| CORE_INSTRUCTION | 1.0 | No |
| CONSTRAINT | 0.9 | Partially |
| OUTPUT_FORMAT | 0.8 | Partially |
| EXAMPLE | 0.5 | Yes |
| EXPLANATION | 0.3 | Yes |
| VERBOSE_DETAIL | 0.2 | Yes (first) |
Protected patterns (boost to 0.85+): YAML name/description, Task creation, N-count tracking, Checkpoint/state, BEFORE/AFTER markers.
Evaluate by semantic understanding, NOT metrics.
| Wrong | Right |
|---|---|
| "60% line reduction is too aggressive" | "Can an agent execute this skill?" |
| "Token ratio exceeds target" | "Are triggers and actions preserved?" |
LLM scores 0-100 based on semantic capability, not line/token ratios. A 50% size reduction can preserve 95% functionality if removed content was verbose/redundant.
Append to .learnings/skill-distiller/calibration.jsonl with metrics and expected score.
Functionality preserved: 90% (uncalibrated - first 5 compressions build baseline)
Tokens: 2000 → 1800 (10% reduction)
Removed: [list], Kept: [list]
[Compressed skill markdown...]
| Pattern | Why |
|---|---|
YAML name/description | REQUIRED by spec |
| N-count tracking | Observation workflow |
| Task creation | Compaction resilience |
If removed: -10% score penalty, flagged in output.
Parallel/serial decisions, performance hints, caching guidance. No score penalty.
Storage: .learnings/skill-distiller/calibration.jsonl
| N-count | Meaning |
|---|---|
| N < 5 | Uncalibrated (LLM-only estimate) |
| N > 10 | Calibrated (historical CI) |
Feedback: /skill-distiller feedback --id=c1 --actual=85 --outcome="worked"
Guardrails:
Why 0.95: Capability loss compounds (0.95 x 0.95 = 0.90 at next level).
| Error | Hint |
|---|---|
| No content | Provide SKILL.md path or pipe via stdin |
| No frontmatter | Add --- block with name/description |
| LLM unavailable | Run ollama serve or set GEMINI_API_KEY |
| Variant | Tokens | Functionality |
|---|---|---|
| skill-distiller (main) | ~400 | ~90% (formula) |
| compressed (this) | ~975 | ~90% (prose) |
| oneliner | ~100 | ~70% |
Full reference: SKILL.reference.md (~2,500 tokens, ~90%)
Token counts use 4 chars/token heuristic (+/-20%). Functionality scores are LLM-estimated.