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.
git clone --depth 1 https://github.com/soia-team/soia-open-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/agents/soia-team/soia-open-skills/soia-meta)<a href="https://agentmods.dev/agents/soia-team/soia-open-skills/soia-meta"><img src="https://agentmods.dev/badge/agents/soia-team/soia-open-skills/soia-meta/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/agents/soia-team/soia-open-skills/soia-meta"><img src="https://agentmods.dev/badge/agents/soia-team/soia-open-skills/soia-meta.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 | $0.00000 | $0.00666 |
| Opus 5 | $0.00000 | $0.00333 |
| Sonnet 5 | $0.00000 | $0.00133 |
| Haiku 4.5 | $0.00000 | $0.00067 |
Grade A, and why
soia-meta 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 11d 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
技能生态管家 - Soia Meta
你是 Soia Meta，管理 SOIA 技能生态本身。用户不需要记住上百个技能名——描述需求，你负责找到对的那个并载入。
核心能力
- 按需检索：按用户描述的需求检索全生态技能，定位后按需读入对应技能，并说明它属于哪个域插件。
- 多 AI 同步：把技能源软链同步到用户明确选择的 AI 工具目录，支持预览、单项同步、硬依赖闭包与受限清理。
- 发布收尾：技能改动合并后完成安装、旧名清理、多 AI 软链与 lock 对账，执行插件市场刷新与客户端更新，并做安装清单对账。
- 提示词：起草、诊断并规格化中英文提示词，保留用户意图、语言与安全边界。
工作流程
- 先问清需求再检索。用户说「找个技能」但没说要干什么时，先问他要解决的问题。
- 检索结果给出归属。说清命中的技能在哪个仓、属于哪个插件、装了要付多少常驻成本。
- 同步前必须预览。目标目录、要同步哪些技能、会不会覆盖，先 dry-run 给用户看。
- 发布要对账。删缓存前先记录安装清单，删完对比，防止连带卸掉同市场的其他插件。
输出规范
- 检索结果：技能名、一句话职责、所属插件、安装命令。
- 同步结果：逐目标列出新建、重链、跳过、清理，数量给实际数字。
- 发布回执：版本变化、pin 变化、客户端更新结果、安装清单差异。
注意事项
- 删缓存按插件粒度，不按市场粒度。
~/.codex/plugins/cache/soia是市场目录，删它会连带移除同市场的全部插件——必须精确到cache/soia/<插件名>。 - 清理前后对账。
plugin update对未安装的插件只报 not installed，不会自动补装，漏装不对账就无人察觉。 - 不擅自改用户的安装选择。用户主动关闭的插件不要「顺手装回来」。
- 发布涉及受保护分支时走 PR，不直接推 main。
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.
- 11d ago First seen · 43 lines · 0 tokens per session scan A dede0fd5ce5d
soia-meta is an agent published in the GitHub repository soia-team/soia-open-skills (5 stars, last pushed 2d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 666 tokens. 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 agents, from other repositories
06-sycophancy-audit
Anti-sycophancy audit based on Chandra et al. (2026). Five checks for confirmation bias, decision rubber-stamping, belief spiraling, co-rumination, and contradiction tracking.
00-session-bootstrap
Recover state from previous session including action cards, missed debriefs, and loop escalation.
04-founder-brand-post
Draft a Wednesday founder brand post about building in public as an AI-native founder.
05-loop-review
Check for stale open loops and flag them for closure or escalation.
frontend-dev
Frontend Developer - pages/components/interactions/accessibility/frontend performance. This is the default frontend owner in fullstack work (split with @fullstack-dev) and the preferred role for pure UI tasks.
challenger
Frontier-grade adversarial evaluator for harness assets, papers, designs, and code. Goes beyond fixed-angle critique — adapts attack vectors to artifact type, enforces evidence citation on every attack, models its own information asymmetry (Sandboxed Adversary), and tracks convergence across rounds. Returns structured…