miloco-terminate-task

任务终止 —— 用户主动取消或任务到期时触发，调 `miloco-cli task delete --reason X` 一笔走完 backend（写 task_terminate_log + 删 task，FK CASCADE 同步清 rule / cron / task_record_* 主表/子表），顺序跑返回的 agent_pending 让 openclaw 老通路 external cron 一并下线。

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

77%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The content is highly actionable with a clear, validated multi-step workflow and real feedback loops for a destructive operation. The two weak spots are conciseness (the worked example overlaps the step tables) and progressive disclosure (everything is inline in one long file with no external references).

Suggestions

Trim the 示例 section so it shows only the end-to-end trace without reprinting the response JSON already given in 第二步, or move the full worked example to a reference file.

Split the output-format JSON schema and the full example walkthrough into a referenced file (e.g. EXAMPLES.md) linked from the body, so the overview stays lean.

Condense the 暂停/算了 branch writeups to reference the 第零步 confirmation table instead of restating its semantics.

Dimension	Reasoning	Score
Conciseness	The body is dense and assumes Claude's competence with no concept-explaining, but the 示例 section re-walks the already-tabulated steps and reprints the JSON response, and the 暂停/算了 branches restate the confirmation table — tightening is possible.	2 / 3
Actionability	Fully executable guidance: exact CLI commands, exact JSON request/response shapes, a copy-paste confirmation script, and explicit branch-mapping tables — copy-paste ready.	3 / 3
Workflow Clarity	Clear 0→1→2→3 sequence with explicit validation (验收标准 code=0 / 2001, 404 noop handling) and feedback loops for the destructive op (errors[] 继续不回滚; partial/failed → system.failureAlert → archive_maintenance 兜底再扫).	3 / 3
Progressive Disclosure	Well-sectioned single file, but it is a ~200-line monolith with no references split out (output JSON schema and worked example are inline); with no bundle files and over 50 lines, the simple-skill exception does not apply.	2 / 3
	Total	10 / 12 Passed

Description

85%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific, complete, and distinctive, with explicit trigger guidance and concrete named actions. Its only weakness is trigger phrasing: it uses event/structural language rather than the natural user utterances that would help routing.

Suggestions

Add the natural user phrases (e.g. 「不喝了 / 不记药了 / 取消那个监控」) into the description so trigger_term_quality reflects what users actually say, not just the event type.

Consider a short 「Use when …」 clause naming the trigger_source values (user_cancel / termination_schedule / archive_maintenance) to make routing even more explicit.

Dimension	Reasoning	Score
Specificity	Names multiple concrete actions — 「调 miloco-cli task delete --reason X」「写 task_terminate_log + 删 task」「FK CASCADE 同步清 rule / cron / task_record_*」「顺序跑返回的 agent_pending … external cron 一并下线」 — matching the multi-action anchor.	3 / 3
Completeness	Explicitly answers both what (the CLI call + transaction cleanup + agent_pending run) and when (「用户主动取消或任务到期时触发」), satisfying the explicit-trigger requirement that would otherwise cap at 2.	3 / 3
Trigger Term Quality	Trigger conditions are present (「用户主动取消或任务到期时触发」) but framed as structural events rather than the natural phrases users say; the natural terms (「不喝了 / 取消那个监控」) live in the body, not the description, so common variations are missing.	2 / 3
Distinctiveness Conflict Risk	A clear niche — a specific CLI command with FK CASCADE backend cleanup and external-cron teardown — with distinct triggers unlikely to collide with other skills.	3 / 3
	Total	11 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_field	'metadata' should map string keys to string values	Warning

	Total	15 / 16 Passed

Repository: XiaoMi/xiaomi-miloco
Path: plugins/skills/miloco-terminate-task/SKILL.md
Commit: f156d34

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.