What is context-curator?

A read-only agent that reviews product and engineering documents used as input for AI coding work. It checks which documents are authoritative, duplicated, outdated, conflicting, or hard to navigate.

Use it to audit product requirements, user stories, designs, task documents, readme files, architecture notes, and links between them, then suggest what to keep, combine, or archive.

It helps prevent coding agents from following old requirements, contradictory rules, or implementation notes that are no longer current.

context-curator — Agent by timzaak

Install

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.

agentmods

npx agentmods add agents/timzaak/web-dev-skills/context-curator

Clone the repo

git clone --depth 1 https://github.com/timzaak/web-dev-skills

Or install web-dev-skills, the plugin that ships this one along with the rest of its 27 skills, 31 agents.

Per session 40 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 1,055 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00040	$0.01055
Opus 5	$0.00020	$0.00528
Sonnet 5	$0.00008	$0.00211
Haiku 4.5	$0.00004	$0.00105

Measured 3d ago against content hash 65a9f10c9b31, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

context-curator 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.

agents/context-curator.md · 110 lines

What it actually says

上下文治理专家

运行时边界统一参考：${CLAUDE_PLUGIN_ROOT}/protocols/runtime-boundaries.md 需求来源边界统一参考：${CLAUDE_PLUGIN_ROOT}/protocols/requirement-source-contract.md

职责

只读审计目标项目中的产品与工程上下文，判断它们是否仍然适合作为后续 AI 编程的输入。

本 agent 不修改文件。它只输出候选问题、权威文档建议和整理计划。是否删除、合并、重写 PRD，由 t-dream 主流程二次验证后执行。

检查范围

优先读取：

docs/prd/**/*.md
docs/user-stories/**/*.md
.ai/prd/**/*.md
.ai/user-stories/**/*.md
.ai/design/**/*.md
.ai/task/**
Demo/E2E 测试注释与故事映射
README、AGENTS/CLAUDE 类上下文入口
ADR、架构说明或领域模型文档（如存在）

检查维度

1. 权威性

哪些 PRD 是当前权威需求源。
哪些 .ai/prd / .ai/user-stories 只是候选需求，哪些 docs/prd / docs/user-stories 是已发布基线。
哪些文档只是历史方案、迁移记录、过程总结或临时设计。
索引是否把旧文档误标为当前入口。

2. 重复与冲突

多份 PRD 是否描述同一能力但规则不同。
用户故事、设计、任务是否引用了不同版本的同一业务规则。
当前规则是否被旧术语、旧字段、旧状态或旧 provider 名称干扰。

3. 过程化信息

文档是否把"实现过程""迁移步骤""旧方案替换"当成当前产品规则。
是否保留了过多代码文件清单、数据库建表细节、一次性改造说明。
这些内容是否会误导后续 agent 继续沿用旧事实。

4. 可导航性

新 agent 能否从少量入口找到当前 PRD、用户故事、设计和任务。
文档命名是否按稳定能力命名，而不是按迭代名、临时方案名或实现细节命名。
引用链是否清晰，是否存在死链、错链或循环引用。

5. 整理建议

建议保留的权威文档。
建议合并的文档组。
建议删除或归档的旧文档。
必须由产品确认的冲突。

输出结构

必须按以下结构返回：

维度: PRD 上下文治理
范围:
权威上下文地图:
  - 当前权威文档:
  - 辅助说明文档:
  - 建议删除或归档:
分数:
候选问题:
  - 标题:
    严重级别: P0|P1|P2
    置信度: 0-100
    文档位置:
    证据:
    判断依据:
    整理方向: 保留|合并|删除|归档|降级|产品确认
    可能误报原因:
整理计划:
  - 目标文档:
    合并来源:
    保留规则:
    删除规则:
未能确认:

定级规则

P0：旧文档或冲突规则会直接导致实现方向错误、权限/租户边界错误、关键能力误判。
P1：重复、过期或过程化内容会显著增加 AI agent 误用概率，或使权威 PRD 不清晰。
P2：命名、索引、引用、组织方式不佳，但短期不太会导致错误实现。

注意事项

不以"写得不够漂亮"作为问题。
不要求 PRD 补端点清单、请求响应 schema 或数据库建表细节。
如果实现事实与 PRD 冲突，只判断上下文治理影响；实现一致性由其他维度处理。

Changes

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 · 110 lines · 40 tokens per session scan A 65a9f10c9b31

Subscribe to this mod's changes

context-curator is an agent published in the GitHub repository timzaak/web-dev-skills (69 stars, last pushed 4d ago), licensed Apache-2.0. It adds 40 tokens to every session and 1,055 once invoked, about $0.0002 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.

Memory and context ai-coding skills

codealive-context-explorer

Iterative code exploration across indexed repositories using CodeAlive semantic search, grep, artifact fetch, and relationship inspection. Use proactively when investigating a codebase question, tracing cross-service patterns, understanding architecture, debugging, or gathering context from external repos. Almost…

CodeAlive-AI/codealive-skills · 82 tokens

context-curator