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 jxoesneon/Ciel --skill rules-distillgit clone --depth 1 https://github.com/jxoesneon/CielWrote 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/jxoesneon/ciel/rules-distill)<a href="https://agentmods.dev/skills/jxoesneon/ciel/rules-distill"><img src="https://agentmods.dev/badge/skills/jxoesneon/ciel/rules-distill/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/jxoesneon/ciel/rules-distill"><img src="https://agentmods.dev/badge/skills/jxoesneon/ciel/rules-distill.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 | $0.00023 | $0.00571 |
| Opus 5 | $0.00012 | $0.00285 |
| Sonnet 5 | $0.00005 | $0.00114 |
| Haiku 4.5 | $0.00002 | $0.00057 |
Grade A, and why
rules-distill 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 9d 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.
What it actually says
CIEL ADAPTATION: Rules Distill (Principle Extraction)
This skill formalizes the mechanism by which CIEL evolves its global ruleset. It applies the "Deterministic Collection + LLM Judgment" principle to identify cross-cutting behaviors that should be promoted from individual skills to mandatory rules.
The Distillation Workflow
Phase 1: Inventory (Deterministic)
The system exhaustively scans the installed skill registry (Ciel/skills/) and the current ruleset to create a baseline inventory of principles.
Phase 2: Cross-Read & Match (Judgment)
An Auditor sub-agent cross-references all skills to find principles that meet the Promotion Criteria:
- Ubiquity: The principle appears in 2+ distinct skills.
- Actionability: It can be written as a "Do X" or "Don't do Y" instruction.
- Risk: There is a clear violation risk if the principle is ignored.
- Novelty: It is not already sufficiently covered in the existing rules.
Phase 3: Review & Execution
Proposed rule updates (Append, Revise, or New File) are presented to the user for explicit approval. Manual approval is mandatory; CIEL never modifies its own ruleset automatically.
Output Format
Each proposal must include:
- Principle: The exact rule text.
- Evidence: List of skills where the principle was observed.
- Violation Risk: What happens if this rule is broken.
- Verdict: Append | Revise | New Section | New File.
Anti-Patterns
- Over-Abstraction: Creating vague rules (e.g., "Code should be clean") that are not actionable.
- Skill Leakage: Moving language-specific hacks that belong in a skill into the global ruleset.
- Automated Refactoring: Modifying rules without a user-in-the-loop review.
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.
- 9d ago First seen · 60 lines · 23 tokens per session scan A d45a072c632e
rules-distill is a skill published in the GitHub repository jxoesneon/Ciel (1 stars, last pushed 5d ago), licensed Apache-2.0. It adds 23 tokens to every session and 571 once invoked, about $0.0001 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 2026-09-03.
Other skills, from other repositories
surge
Use when a user provides a PRD, spec, or detailed requirements document and needs a full project delivered through iterative expert orchestration — multi-round analyze/research/design/implement/QA cycles with convergence detection. NOT for: single-file edits, quick prototypes, simple Q&A, or tasks without a written…
ai-native-software-dev
Use when the user asks to build, change, modernize, or reconstruct non-trivial software with an AI-native workflow, including greenfield development, work in an existing codebase, or authorized reference-based product reconstruction; also use when lifecycle planning, architecture, implementation, testing, review…
Code Quality Review
Conduct a comprehensive code quality review focused on clean code, architecture, and project standards.
redpolicy-guide
An editorial policy for turning webinar transcripts into self-contained guides. A transcript is a written record of spoken material, while a self-contained guide can be understood without watching the webinar.
write-overview
A workflow for researching and writing a review essay about a project topic when no finished source text exists. It gathers the facts, studies example essays when available and produces the article in a project output folder.
redpolicy-article
An editorial guide for writing comparison articles about services, tools, and products. It defines a consistent entry format covering the name, link, price, description, and notable features.