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/hugorcd/evlog/content-passnpx skills add HugoRCD/evlog --skill content-passgit clone --depth 1 https://github.com/HugoRCD/evlogWhat 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 | $0.00154 | $0.03175 |
| Opus 5 | $0.00077 | $0.01588 |
| Sonnet 5 | $0.00031 | $0.00635 |
| Haiku 4.5 | $0.00015 | $0.00317 |
Grade A, and why
content-pass 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 3d 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 — 175 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Content pass
One pass, one group, one pull request. The corpus is ~120 files: the docs tree and the landing, the four package READMEs, the internal and published skills, and the three AGENTS.md files. None of that gets fixed in a day, and trying is how a rewriter starts rewriting for its own sake.
Half the corpus is read by people and half by agents, and the pass treats them differently. A skill or an AGENTS.md governs the agent running this pass, so it may fix a house rule there (punctuation, a dead link, a retired entry point, a wrong term) and nothing else. Procedure, bounds, and a skill's description come back as findings for Hugo. That is M-09 in the doctrine, and content__targets enforces it by returning those files with mode report.
The doctrine lives in the repository, at .agents/skills/write-evlog-content/. This file is the procedure; that skill is the standard. Never restate its rules here. Read them there, and when they are wrong, fix them there.
The two halves
Rewrite is the default: the scanner ranks the corpus, the worst pages get reviewed, and what survives review gets applied.
Enrich is what runs when the rewrite half comes back empty, with no page above the bar or everything ranked inside its cooldown. That is a good day, not a wasted one. Switch to the second half rather than lowering the bar.
Never run both in one pass. A PR that rewrites two pages and adds a third is a PR nobody reviews properly.
Rewrite
1. Pick the targets
Call content__targets. It runs the scanner over the whole corpus, drops files changed inside the cooldown, and returns the top files from a single group with their candidates. A group is one docs section, one skill's directory, or a flat surface (readme, agents, landing).
Pass surface when Hugo asked for one, or when the weekly corpus check found a surface drifting. Otherwise take what ranks.
Read eligible before deciding anything. It is the count of files with findings that are outside the cooldown, and it is the only number that says whether there is work. eligible: 0 sends you to Enrich. Anything above zero means the rewrite half has targets, and the pass rewrites.
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.
- 3d ago First seen · 175 lines · 154 tokens per session scan A 72ca7911e505
content-pass is a skill published in the GitHub repository HugoRCD/evlog (1,831 stars, last pushed yesterday), licensed MIT. It adds 154 tokens to every session and 3,175 once invoked, about $0.0008 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 2026-08-30.
Other skills, from other repositories
frontmcp-observability
Use when adding tracing, structured logging, metrics, or monitoring to a FrontMCP server. Covers zero-config OpenTelemetry distributed tracing across all flows; the this.telemetry API for custom spans, events, and attributes in tools, plugins, agents, and skills; structured JSON logging with trace correlation and…
golang-observability-opentelemetry
Instrumenting Go applications with OpenTelemetry for distributed tracing, Prometheus for metrics, and structured logging with slog.
Observability Checklist
Reviews a service or codebase against a full observability checklist — logs, metrics, traces, and alerting gaps.
anyrobot-shadcn-frontend-js
AnyRobot 前端开发核心指南：提供 JavaScript + React + shadcn/ui 开发的精简实用规范和最佳实践。 当用户需要前端开发指导、组件开发或项目结构建议时触发。 本技能提供实际可用于项目开发的核心代码示例和开发指南。.
数据模型文档生成
当用户要生成或更新 Doris 表的数据模型 Markdown（三一级标题：数据摘要、数据表列、JSON扩展字段）、从 Doris 拉取元数据与至少 100 行样本，并由脚本或 --llm 补全说明时启用。.
fix-loop
Use when a Java app that has stacktale installed is failing — after running it or its tests, to find what broke, fix it, and confirm the fix by re-running and checking for new errors. Also use when asked to "check the errors", "what's failing", or to work through errors-ai.log.