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 riwonswain-ovo/OfferLoop --skill job-collectiongit clone --depth 1 https://github.com/riwonswain-ovo/OfferLoopWrote 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/riwonswain-ovo/offerloop/job-collection)<a href="https://agentmods.dev/skills/riwonswain-ovo/offerloop/job-collection"><img src="https://agentmods.dev/badge/skills/riwonswain-ovo/offerloop/job-collection/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/riwonswain-ovo/offerloop/job-collection"><img src="https://agentmods.dev/badge/skills/riwonswain-ovo/offerloop/job-collection.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 | $0.00077 | $0.04113 |
| Opus 5 | $0.00039 | $0.02056 |
| Sonnet 5 | $0.00015 | $0.00823 |
| Haiku 4.5 | $0.00008 | $0.00411 |
Grade A, and why
job-collection 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 8d 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 — 201 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Job Collection
把用户有权访问的结构化招聘信息源同步到既有的「求职企业清单」，并保持筛选、去重、待确认写入、通知和失败恢复可追溯。
运行脚本时以本文件所在目录为 Skill 根目录；scripts/... 和 references/... 均相对该目录解析。
触发边界（选择 Skill 前）
- 用户提供受支持的招聘表格链接、要求同步已登记来源、处理待确认候选或查询既有清单时触发。
- 用户只要求去互联网或其他平台寻找岗位，且没有受支持的来源链接或已登记来源上下文时不触发；不得调用 来源扫描或目标写入工具，只说明本 Skill 仅处理用户提供或已经登记的飞书、腾讯表格来源。
- 只发一个无法确认用途的飞书表格链接时，不猜测它属于 OfferLoop；先用一句话确认是否要筛选整理其中岗位。
执行硬门禁（先执行）
完整登记本轮来源记录后，严格按 record.normalize → 每条记录 candidate.route → 适用记录
candidate.dedupe → auto_write 的 target.write → target.verify / awaiting_write_confirmation 的
pending.create → 通知 → 游标 → evaluation.finalize 执行。所有记录都完成标准化和独立路由后才能去重；
所有记录取得可验证处置后才能通知、推进游标或 finalize。第一次 finalize 就必须成功，不得用
E_FINAL_INCOMPLETE 或其他拒绝结果发现遗漏后补调用。
每条记录使用 stable_key=source_id:source_record_id。单记录调用顶层传
stable_key、source_id、source_record_id，candidate.route 还必须传精确 route；批量调用顶层传
非空、去重的 stable_keys，并让 records 中每项携带相同三个身份字段和精确 route。禁止只传城市、
年份、pending 或无法追溯的嵌套对象。route 只允许 hard_filtered、
awaiting_write_confirmation、auto_write。
一个真实链接足以保持原岗位 route；公告和投递链接都缺失时直接返回 hard_filtered，不写入、不去重、
不创建待确认编号或借用相邻链接。该门禁优先于不确定筛选条件。岗位硬排除只在
job_scope_complete=true 且完整范围内全部岗位明确排除时成立；范围不完整必须待确认。多个独立岗位同时
含排除与明确考虑方向时只保留考虑岗位并 auto_write；只有同一岗位语义冲突才待确认。完成摘要只能由
工具确认的写入回读、待确认持久化和明确处置计数生成。
使用场景
- 登记来源：只读探测用户提供的招聘信息源并保存定位信息。
- 增量同步：读取已登记的活跃来源，筛选、去重并写入新岗位。
- 待确认写入：按稳定编号处理需要用户决定的候选。
- 清单查询：按
待确认、感兴趣、已投递或已拒绝查询企业清单；不读取招聘来源。
支持的来源：
- 飞书/Lark 多维表格：
feishu.cn/base/或larksuite.com/base/。 - 腾讯 Smartsheet：
docs.qq.com/smartsheet/，只通过已配置并验证的官方 MCP 读取。
开始前检查
本 Skill 的第一项动作是读取 ../.offerloop-runtime/references/installation-mode.md 并运行模式检查。
随后必须运行 python3 scripts/notification_authorization.py check。返回 authorized=true 时，表示用户已对
输出中的准确群名和发送身份授予 job-collection 长期通知授权：本轮以及后续运行完成后直接发送，不得在
同步末尾再次询问“是否发送”或“确认发送”。通用预检把飞书通知标为 unverified 时，以本脚本对当前
目的地的绑定结果为准；工具或操作系统仍可展示自身必要的权限提示。
返回 authorized=false 时，在任何同步和外发之前展示脚本返回的准确群名与身份，只询问一次是否为
job-collection 开启长期自动发送。用户同意后运行
python3 scripts/notification_authorization.py authorize --confirm-standing-authorization，再开始同步。默认群、
目标 ID 或发送身份发生变化会使旧授权自动失效，必须为新目标重新取得一次授权。不得把一次授权扩展到
其他 Skill、其他群或其他消息类型。
What ships with it
31 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/dedup_judge.md 4.1 KB
- references/excel-insert.md 4.3 KB
- references/failure-handling.md 7.0 KB
- references/field-contract.md 9.6 KB
- references/init-workflow.md 5.9 KB
- references/notification.md 8.2 KB
- references/personal-excel-source.md 7.7 KB
- references/prewrite-confirmation.md 10 KB
- references/tencent-smartsheet-source.md 7.1 KB
- scripts/dedupe_candidates.py 17 KB runs code
- scripts/execution_contract.py 15 KB runs code
- scripts/incremental_scan.py 9.0 KB runs code
- scripts/notification_authorization.py 7.3 KB runs code
- scripts/preference_migration.py 2.9 KB runs code
- scripts/progress_sync.py 5.9 KB runs code
- scripts/sync_pipeline.py 49 KB runs code
- scripts/sync_utils.py 15 KB runs code
- scripts/tencent_mcp.py 27 KB runs code
- scripts/tencent_mcporter.py 6.9 KB runs code
- scripts/validate_skill.py 11 KB runs code
- tests/test_dedupe_candidates.py 5.8 KB runs code
- tests/test_execution_contract.py 8.4 KB runs code
- tests/test_incremental_scan.py 7.5 KB runs code
- tests/test_notification_authorization.py 2.2 KB runs code
- tests/test_preference_migration.py 2.0 KB runs code
- tests/test_progress_sync.py 13 KB runs code
- tests/test_sync_pipeline_missing_links.py 2.5 KB runs code
- tests/test_sync_pipeline.py 26 KB runs code
- tests/test_sync_utils.py 14 KB runs code
- tests/test_tencent_mcp.py 14 KB runs code
- tests/test_tencent_mcporter.py 2.8 KB runs code
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.
- 8d ago Changed · +24 lines 607dc30c175e
- 12d ago First seen · 177 lines · 77 tokens per session scan A b430130be394
job-collection is a skill published in the GitHub repository riwonswain-ovo/OfferLoop (16 stars, last pushed 6d ago), licensed MIT. It adds 77 tokens to every session and 4,113 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 2026-08-30.
Other skills, from other repositories
linear-cli
Use when the user wants to list, view, start, create, or update Linear issues from the terminal via schpet/linear-cli, including setup, auth, repo config, and safe read/write boundaries.
feedback-workflow
A command-line workflow for reading and responding to approved software feedback. It keeps track of the feedback version, approval, work claim, evidence, and result.
discussion-participant
A command-line workflow for participating in private discussions and replying through NextClaw. It verifies participant identity from trusted platform data rather than from message text.
lean-in
Applies Sandberg's Lean In framework to diagnose and address gender barriers in organizations, covering the ambition gap, mentor-sponsor deficit, and structural inclusion deficits. Use when designing inclusive hiring processes, building sponsorship programmes, analyzing gender diversity gaps, retaining senior women…
feishu-task
A Feishu task-management tool for creating and organizing to-do items and task lists. Feishu is a workplace collaboration app.
lark-task
A Feishu task-management tool for handling to-do items, task lists, assignments, subtasks, attachments, and task-focused agents. Feishu Tasks is Feishu's workspace for tracking work and responsibilities.