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 SimpleEve/eo-skills --skill eo-recallgit clone --depth 1 https://github.com/SimpleEve/eo-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/simpleeve/eo-skills/eo-recall)<a href="https://agentmods.dev/skills/simpleeve/eo-skills/eo-recall"><img src="https://agentmods.dev/badge/skills/simpleeve/eo-skills/eo-recall/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/simpleeve/eo-skills/eo-recall"><img src="https://agentmods.dev/badge/skills/simpleeve/eo-skills/eo-recall.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00088 | $0.01428 |
| Opus 5 | $0.00044 | $0.00714 |
| Sonnet 5 | $0.00018 | $0.00286 |
| Haiku 4.5 | $0.00009 | $0.00143 |
Grade A, and why
eo-recall 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 6d 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 — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
eo-recall — 回忆与解释
项目记忆与代码的只读检索入口：把「我忘了当初……」变成一次有出处、分口径的检索问答。严格只读——除 tmp/eo/explain/ 的一次性解释页外不写任何文件。
核心原则
- 查了才答：答案必须来自本次检索到的文档/代码，每个论断带出处（文档路径或 file:line）
- 无记载不编造：文档里查不到就明说「文档无记载，以下是我从代码现读的」——把「回忆」和「现场推断」显式分开
- 口径分层：需求设计视角与代码实现视角分开作答，不搅在一起
- 定位纪律：
.codegraph/索引存在则codegraph explore优先；不存在则走 INDEX 收敛 + 源码直读，不全局盲扫
前置
必须能找到 .eo-project.json。同目录存在 .eo-project.local.json 时顶层字段覆盖合并（local 优先）。找不到 → 报错退出，提示运行 /eo-project-init。
工作流程
第一步：问题分类（可复合）
| 问的是 | 类型 | 主信源 |
|---|---|---|
| 「现在是什么样 / 规则是什么」 | 现状 | state/ 现状篇（state.enabled 时）→ codegraph（有索引时）→ 源码 |
| 「代码在哪、怎么实现的」 | 实现 | codegraph（有索引时）→ 源码 |
| 「当时为什么这么定」 | 缘由 | change §1 已钉决策 → decisions/ → brainstorm/ |
复合问题（最常见：「这个系统当时怎么设计的」= 现状 + 缘由）按需叠加信源。
第二步：检索瀑布
- 现状/实现：
state/已启用（配置state.enabled）时先读eo-doc/state/对应现状篇（派生快照；篇≠码以代码为准，并提示可跑/eo-doc-manager sync刷新）；随后.codegraph/索引存在则codegraph explore "<自然语言问题>"优先召回相关符号源码与调用链；不存在则按目录结构收敛、直读源码相关段落（不通读文件） - 缘由：
eo-doc/changes/INDEX.md找相关 change → 读其 §1 意图与已钉决策；再查<project_root>/decisions/（扫其 INDEX.md，无 INDEX 则按文件名收敛）与<project_root>/brainstorm/的相关记录；涉及外部世界（第三方 API / 平台规则 / 竞品 / 选型）→ 按 ../eo-shared/research.md 消费规则查<project_root>/research/；lessons 若 trigger 命中也一并带上（踩坑背景常常就是「为什么后来改成这样」） - 候选 >3 或全无命中 → 追问关键词，不硬猜
第三步：分层作答
## 需求上（设计口径）
<规则/流程/边界，出处：change <id> §1、decisions/2026-xx.md>
## 代码上（实现口径）
<入口、关键路径、数据流，出处：src/xx.ts:120-160>
## 当时为什么这么定（若问及）
<已钉决策 + 理由，出处>
⚠️ 文档与代码不一致处 / 文档无记载处，单独列出（以代码为准）
- 不确定的就说不确定，不填补空白
第四步：富输出（按需，默认不产文件）
- 流程 / 状态机 / 多角色交互类逻辑 → 主动提议：「画个 mermaid？」（对话内直接给，规范见 ../eo-doc-manager/references/mermaid.md）
- 盘根错节的大逻辑（多系统联动、你需要反复对照的）→ 提议生成一次性自包含 HTML 解释页：
tmp/eo/explain/<date>-<topic>.html，零外部依赖、用项目真实名词与数据、可交互（点击展开/分步演示）；它是可丢弃工件（见 ../eo-shared/conventions.md），有长期价值的结论应沉淀到 decisions/lessons（提示用户走 /eo-project-record）而不是留在 tmp
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.
- 6d ago First seen · 76 lines · 88 tokens per session scan A f53c75204bc0
eo-recall is a skill published in the GitHub repository SimpleEve/eo-skills (31 stars, last pushed 3d ago), licensed MIT. It adds 88 tokens to every session and 1,428 once invoked, about $0.0004 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 2026-09-04.
Other skills, from other repositories
close-my-loops
Build and maintain a personal work profile and open-loop tracker for the user's own job — a markdown file covering their org, key contacts, communication style, active initiatives, and open loops synthesized from their own email/Slack; a Cowork Project setup so that file is actually read every session; a scheduled…
project-compass
Map the tribal knowledge of a codebase into per-area COMPASS.md files — the why, the gotchas, and how each area connects — for existing repos and as they grow. Use when the user says "map this codebase", "generate compass files", "refresh the compass", or after a significant area changes.
aisee:reflect
A guide for reviewing a work session and saving reusable lessons as project documents. It distinguishes temporary suggestions from confirmed long-term team knowledge and from records of specific solved problems.
aisee:knowledge-curate
A review process for turning project notes into reusable team-knowledge drafts. It checks candidate lessons for private details, duplicates, outdated information, and usefulness beyond one project.
failure-mode-capture
Record a "don't do X here, it breaks Y" lesson into AGENTS.md so an agent doesn't repeat a mistake. Dedupes against CLAUDE.md/instincts before writing, promotes generalizing lessons out of memory, and keeps AGENTS.md under budget. Use after a bug, regression, or near-miss, or when the user says "capture this" / "make…
handoff
Carry essential working context to another task, directory or worktree, harness, collaborator, or mid-phase side fork in a portable Markdown handoff.