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 ryanzhao1011/workframe --skill maintenance-reviewgit clone --depth 1 https://github.com/ryanzhao1011/workframeWrote 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/ryanzhao1011/workframe/maintenance-review)<a href="https://agentmods.dev/skills/ryanzhao1011/workframe/maintenance-review"><img src="https://agentmods.dev/badge/skills/ryanzhao1011/workframe/maintenance-review.svg" alt="Measured on agentmods" 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 | $0.00026 | $0.01428 |
| Opus 5 | $0.00013 | $0.00714 |
| Sonnet 5 | $0.00005 | $0.00286 |
| Haiku 4.5 | $0.00003 | $0.00143 |
Grade B, and why
maintenance-review scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
python "$(cat .claude/workframe-state/plugin-root.txt)/scripts/maintenance_workorder.py" \ How it starts
The opening of the file, as written. The whole thing — 94 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/core:maintenance-review 唤醒后手动维护
用途
dormant 项目被重新打开后，自动维护链路全部暂停。用户主动 /core:maintenance-review 才恢复正常维护。
Claude 不会自动调用（disable-model-invocation: true）。
输入
两种调用模式：
- 完整 wake-up review（无参数）：按顺序走完 3 个检查点，每步都用
AskUserQuestion确认是否继续 - 快速 dismiss 模式（
--dismiss <PM-ID>）：把指定 ID 的pending_maintenance条目关闭（经模式 A 的代码通道执行；audit 保持只读，写操作归本 skill）。不触发 librarian / self-iteration
执行步骤
模式 A：--dismiss <PM-ID>（快速关闭单条）
-
Read
.claude/workframe-state/activity-state.json，确认id == <PM-ID>的条目存在且status == "open"，记下它的kind/severity（用于第 3 步向用户复述） -
用代码通道关闭（不要自己改这个文件）——它一并写
pending_maintenance_dismissed事件（ts/at由脚本填同一时刻，二者都是 schema 必填）：python "$(cat .claude/workframe-state/plugin-root.txt)/scripts/maintenance_workorder.py" \ --close-pm <PM-ID>activity-state.json还装着session_counter/recent_drift_repairs/ 注入标记等 攒出来的状态，整份重写漏字段不报错、只静默丢历史；脚本走文件锁 + 原子替换 + 三方合并， 只动pending_maintenance一处。 -
在响应中输出：
已关闭 PM-ID=<PM-ID>，原 kind=<X> severity=<Y> -
不触发 librarian / self-iteration；该条目会在 SessionStart 的 7 天 GC 后从数组中清理
模式 B：完整 wake-up review（无参数）
-
会话摘要
- Read
.claude/workframe-state/activity-state.json— 告知 dormant 多少天 - Read
.claude/workframe-state/session-digest-latest.md— 上次活跃时的摘要 - 问用户："继续执行 Librarian 整理吗？" (A: 是 / B: 跳过本步 / C: 取消全部)
- Read
-
Librarian 整理（如用户同意）
- 以
manual-maintenance=true上下文调用 librarian skill - 用户已确认继续整理时，允许处理 dormant / wake_up_pending 下被冻结的 promotion 候选和 decay / 容量候选
- 输出整理摘要（整理了几条、提升几条、确认降级几条、仍待审查几条）
- 问："继续执行 self-iteration 吗？"
- 以
-
Self-iteration 提案（如用户同意）
- 调用 self-iteration skill
- 输出提案摘要（生成几个候选、阶段 4 审批清单）
- 问："提案是否批准执行？（逐条选）"
-
结尾
-
关闭本次处理过的 pending_maintenance 条目——走模式 A 第 2 步的
--close-pm <PM-ID> ...（可一次传多个 ID），不要手改 activity-state.json -
结束 wake-up 状态（
dormant=false/wake_up_pending=false）同样走代码通道：python "$(cat .claude/workframe-state/plugin-root.txt)/scripts/maintenance_workorder.py" --wake-done
-
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.
- 7d ago First seen · 94 lines · 26 tokens per session scan B 3695ea689eda
maintenance-review is a skill published in the GitHub repository ryanzhao1011/workframe (4 stars, last pushed 20d ago), licensed MIT. It adds 26 tokens to every session and 1,428 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
html-prototype
A method for creating single-file HTML mockups that open in a browser and follow a project's design tokens, such as its fonts, colours, spacing, corners, and shadows.
feature-acceptance
A browser-based acceptance testing process that compares an online implementation with a product requirements document and, when available, an HTML prototype. It records what is complete, partial, missing, or inconsistent, with screenshots and prioritized blockers.
prd
A skill for writing and reviewing product requirements documents, or PRDs: documents that describe what a product or feature should do. It covers structure, fields, rules, edge cases, metrics, page layouts, and related system or algorithm flows.
gbg
Convert any plan, PRD, feature idea, or brief into a sequenced goal-by-goal execution document with reviewer-gated commits — for everyday feature development, MVPs, migrations, refactors, and hardening alike. Generates a project brief, per-goal scope/tasks/acceptance criteria, an LLM-reviewer prompt (Codex/etc.), a…
ui-walkthrough
A UI comparison skill that checks a design HTML file against a live webpage at the same screen size and state. It reads computed styles, the actual values applied by the browser, to document differences in layout, components, colors, fonts, corners, icons, and spacing.
competitor-research
A skill for maintaining documents about competing products and AI research. It covers product features, agent modules, and framework comparisons, with rules for organizing facts, judgments, screenshots, and follow-up items.