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 skills add igapyon/igapyon-agent-skills --skill igapyon-skill-compactorgit clone --depth 1 https://github.com/igapyon/igapyon-agent-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/igapyon/igapyon-agent-skills/igapyon-skill-compactor)<a href="https://agentmods.dev/skills/igapyon/igapyon-agent-skills/igapyon-skill-compactor"><img src="https://agentmods.dev/badge/skills/igapyon/igapyon-agent-skills/igapyon-skill-compactor/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/igapyon/igapyon-agent-skills/igapyon-skill-compactor"><img src="https://agentmods.dev/badge/skills/igapyon/igapyon-agent-skills/igapyon-skill-compactor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00116 | $0.01475 |
| Opus 5 | $0.00058 | $0.00737 |
| Sonnet 5 | $0.00023 | $0.00295 |
| Haiku 4.5 | $0.00012 | $0.00147 |
Grade A, and why
igapyon-skill-compactor 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 8d ago.
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 — 141 lines — stays where its author put it; the contents beside it link to each section on GitHub.
igapyon-skill-compactor
Compact Agent Skills for repeated runtime efficiency without silently changing activation, behavior, safety, output, or validation contracts.
Core Execution Contract
Follow this loop for every compaction. Do not require a reference file to carry these steps.
- Establish the target, optimization surface, mode, allowed loss, and writable scope. If the user supplied them clearly, proceed without another question.
- Extract the source contract and a typed structured inventory.
- Make a placement map:
keep,move,rewrite,delete,split, orno change. - Change only the authorized target and preserve the selected mode's contract.
- Re-extract the inventory from the result and compare it with the source.
- Repair critical regressions, then report measured change and validation.
The source contract includes:
- activation and non-activation boundaries
- inputs, outputs, required order, conditions, and postconditions
- commands, paths, IDs, URLs, code/config/API examples, and concrete evidence
- prohibitions, exceptions, fallbacks, safety rules, and human confirmations
- validation steps, expected results, references, and output contracts
Treat any missing critical item as a regression unless the user explicitly accepted that loss.
Mode And Loss Boundary
conservativeis the default. Preserve all critical inventory and ambiguous source-specific meaning. Prefer local deduplication or moving long material with a direct route over deletion.structuralapplies only when the user wants stronger compaction while retaining explicit lists, steps, criteria, and operational structure.summaryapplies only when the user requests summary-like compression or accepts lower reconstruction fidelity. Never weaken activation, safety, required validation, or output contracts silently.no changeis valid when the target is already compact or when added references and maintenance cost would outweigh the reduction.
What ships with it
48 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 213 B
- distilled/.gitkeep 1 B
- distilled/token-efficiency-design-essence.md 7.9 KB
- examples/compaction-modes-example.md 4.8 KB
- examples/frontmatter-distilled-example.md 709 B
- examples/frontmatter-reference-example.md 414 B
- index.json 14 KB
- README.md 13 KB
- references/agent-skill/compaction-modes.md 11 KB
- references/agent-skill/compaction-workflow.md 6.6 KB
- references/agent-skill/distilled-materials.md 1.6 KB
- references/agent-skill/frontmatter-templates.md 2.1 KB
- references/checklists/agent-skill.md 9.5 KB
- references/checklists/checklist-timing.md 1.4 KB
- references/checklists/INDEX.md 569 B
- references/checklists/prompt.md 3.4 KB
- references/checklists/workflow.md 12 KB
- references/system/checklist-selection.md 893 B
- references/system/design-philosophy.md 3.5 KB
- references/system/token-efficiency-techniques.md 20 KB
- templates/frontmatter-distilled.md 930 B
- templates/frontmatter-minimal.md 303 B
- tests/activation-prompts.jsonl 1.7 KB
- tests/baseline-known.json 901 B
- tests/behavior-prompts.jsonl 5.4 KB
- tests/evaluator-prompt.md 924 B
- tests/expected-result-schema.json 797 B
- tests/fixtures/branching-flow/expectations.json 454 B
- tests/fixtures/branching-flow/SKILL.md 717 B
- tests/fixtures/code-example/expectations.json 464 B
- tests/fixtures/code-example/SKILL.md 594 B
- tests/fixtures/concrete-data/expectations.json 555 B
- tests/fixtures/concrete-data/SKILL.md 661 B
- tests/fixtures/deploy-helper/expectations.json 1007 B
- tests/fixtures/deploy-helper/SKILL.md 1.1 KB
- tests/fixtures/duplicate-context/expectations.json 495 B
- tests/fixtures/duplicate-context/SKILL.md 698 B
- tests/fixtures/no-change/expectations.json 338 B
- tests/fixtures/no-change/SKILL.md 233 B
- tests/fixtures/prose-embedded/expectations.json 505 B
- tests/fixtures/prose-embedded/SKILL.md 623 B
- tests/fixtures/writing-contract/expectations.json 526 B
- tests/fixtures/writing-contract/SKILL.md 611 B
- tests/harness-lib.mjs 8.6 KB runs code
- tests/harness-lib.test.mjs 6.1 KB runs code
- tests/INDEX.md 4.0 KB
- tests/reference-routing-prompts.jsonl 2.3 KB
- tests/run-codex-prompt-tests.mjs 17 KB runs code
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.
- 8d ago First seen · 141 lines · 116 tokens per session scan A b8674992d216
igapyon-skill-compactor is a skill published in the GitHub repository igapyon/igapyon-agent-skills (2 stars, last pushed 24d ago), licensed Apache-2.0. It adds 116 tokens to every session and 1,475 once invoked, about $0.0006 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
release-version
Steps to cut a VNote version release (e.g. v4.3.0) — bump the version with scripts/updateversion.py, refresh and fill zhCN/ja translations via lupdate, write the changes.md changelog, and trigger the CI [Release] build. Use when asked to "release", "cut a release", "bump the version", or "prepare vX.Y.Z".
clarity
Draft, rewrite, or review reader-facing prose so it is specific, useful, and recognizably the author's without inventing facts or performing humanness. Use for essays, articles, newsletters, documentation, talks, launch copy, and other important prose that feels generic, hollow, or AI-shaped. Supports co-write…
novel-creator
A Chinese-language writing workflow for creating, continuing, reviewing, polishing, and saving novel chapters.
novel-reviewer
A review and editing workflow for existing novel chapters. It checks story logic, timelines, character states, unresolved clues, world rules, and writing style.
dialoguequality
A dialogue-review system for novels that studies how characters speak and whether their voices remain distinct. It can examine speech habits, sentence length, question frequency, and repeated phrases.
foreshadowing-system
A system for tracking foreshadowing in novels. Foreshadowing means placing a clue or unresolved question earlier in a story and addressing it later.