Borrowing it
Nothing to install: this file belongs to nirholas/three.ws. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/nirholas/three.ws/main/.agents/skills/okx-task-watch/SKILL.mdgit clone --depth 1 https://github.com/nirholas/three.wsWrote 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/nirholas/three.ws/okx-task-watch)<a href="https://agentmods.dev/skills/nirholas/three.ws/okx-task-watch"><img src="https://agentmods.dev/badge/skills/nirholas/three.ws/okx-task-watch/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/nirholas/three.ws/okx-task-watch"><img src="https://agentmods.dev/badge/skills/nirholas/three.ws/okx-task-watch.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 70 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00138 | $0.05983 |
| Opus 5.5 | $0.00055 | $0.02393 |
| Sonnet 5.5 | $0.00028 | $0.01197 |
| Haiku 4.5 | $0.00014 | $0.00598 |
Grade A, and why
okx-task-watch 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 2mo 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 — 217 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OKX Task Watch
Live monitor for the user-session task inbox. Owns: triggers, the watch command, anti-cron rules, item dispatch (notification / decision_request), claim semantics, llmContent execution, stop conditions.
Business actions (apply / deliver / dispute / quote / accept) belong to okx-agent-task. This skill only handles the watch loop.
Triggers — MANDATORY ROUTING (NOT cron / scheduled / polling)
When ANY trigger phrase below matches, execute §Action. The watch command is a single long-poll call (the call itself waits for events); do NOT wrap in /loop, Cron, $CODEX_HOME/automations, watch -n, sleep loops, scheduler frameworks, or "set up a recurring job". Do NOT ask "how often should I check?" — the long-poll IS the wait. Do NOT substitute another command for polling.
Trigger phrases:
- Chinese (live monitor):
监听任务进展/开始监听任务/关注任务进展/使用监听 skill 监听任务进展/帮我盯着任务/任务有进度就告诉我/任务有动静告诉我/开监听/watch 任务 - Chinese (history / backlog drain):
历史消息/历史记录/过去消息/之前的消息/帮我看看之前的历史消息/看下之前的消息/未读消息 - Chinese (continuation — clarify first, see §Continuation triggers):
继续监听/继续盯着/继续 watch/接着监听/再监听一下/继续监听任务 - English (live monitor):
task watch/user watch/monitor task progress/keep me posted on tasks/watch tasks/start watching - English (history / backlog drain):
show past messages/show message history/catch me up on tasks/unread task messages - English (continuation — clarify first, see §Continuation triggers):
keep watching/continue watching/resume monitoring
⚠️ Continuation triggers are a special case — they do NOT immediately call watch. They imply the user wants to keep watching some specific task, but the intent is ambiguous (which task? or all of them?). See §Continuation triggers below for the clarification flow.
📥 Why "view history" routes here: watch is a destructive read of the event stream — each call returns the full backlog of unread events accumulated since the last call (e.g. while no one was watching), then long-polls for new ones. A user asking for past / missed / unread messages is asking to drain that backlog — same command, same Dispatch flow. Do NOT route to
agent active-tasks/agent status(those are summaries, not the actual notification bodies). For un-replieddecision_requestitems specifically (whichwatchalready consumed but the user hasn'tchecked), see §"Pull outstandingdecision_requestitems".
What ships with it
3 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.
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.
- 2mo ago First seen · 217 lines · 138 tokens per session scan A 242c5613841c
okx-task-watch is a skill published in the GitHub repository nirholas/three.ws (269 stars, last pushed yesterday), licensed Apache-2.0. It adds 138 tokens to every session and 5,983 once invoked, about $0.0006 per session on Opus 5.5. 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.
okx-task-watch compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| okx-task-watchnirholas/three.ws | Skill | 269 | yesterday | Apache-2.0 | A | 138 | Claude Code, Codex |
| cesiumjsznlgis/opengis-skills | Skill | 76 | 10d ago | MIT | A | 93 | Claude Code, Codex |
| decision-avatarellmos-ai/skills | Skill | 9 | 4d ago | MIT | A | 49 | Claude Code, Codex |
| sheleg-designssheleg/sheleg-design-skill | Skill | 4 | 4d ago | MIT | A | 80 | Cursor |
| p5jsmateaix/mateclaw | Skill | 1.2k | yesterday | Apache-2.0 | A | 20 | Claude Code, Codex |
| threejs-docspledgeandgrow/pledge-skills | Skill | 10 | 3mo ago | MIT | A | 112 | Claude Code, Codex |
| 3d-website-architectdeveshpunjabi/3d-website-skill | Skill | 8 | 8mo ago | MIT | A | 240 | Claude Code, Codex |