CtrlK
BlogDocsLog inGet started
Tessl Logo

agent-module-architecture

处理 BK-CI Agent 构建机侧能力时使用,例如守护进程、心跳、Ask 轮询、任务拉起、升级更新和与 Dispatch/Worker 的协作。当用户要改构建机宿主侧行为而不是 Worker 执行细节时优先使用。

65

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Medium

Suggest reviewing before use

SKILL.md
Quality
Evals
Security

Quality

Content

67%Weight 40%Scale 1-5

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

The body is a compact, well-organized router for an architecture-navigation skill: clear scope/anti-scope sections, a problem-type routing map, and dense rules and pitfalls. Its main weaknesses are thin actionability (only document titles, no executable specifics) and reference paths that cannot be verified in the bundle plus duplicated reference listings.

Suggestions

Add one line per reference doc summarizing what it contains (key classes, entry points, or commands) so the routing decision is actionable before the doc is opened.

Remove the duplicate reference list in 延伸阅读 or the routing map in 快速指导 — keeping one clearly signaled navigation section would tighten the body.

Verify the reference paths (reference/1-runtime-upgrade.md etc.) exist in the skill bundle, or fix the directory name (reference/ vs references/) so the pointers resolve.

Add a fallback rule for problems that span both Agent and Worker (e.g., which doc to read first) to close the workflow gap.

DimensionReasoningScore

Conciseness

The body is lean and assumes competence — sections like 高信号规则 and 关键陷阱 deliver dense, non-obvious guidance with no padding about concepts Claude already knows. It falls short of anchor 5 because the three reference docs and the worker/dispatch handoffs are each stated twice (快速指导 steps 3–5 and again in 延伸阅读), a small redundancy that could be trimmed.

4 / 5

Actionability

There is a concrete problem-type-to-document routing map ("Ask、心跳与任务拉起:reference/2-ask-heartbeat-build.md") and explicit handoff rules, but beyond document titles there is no executable detail — no commands, source paths, class names, or even a summary of what each reference covers. This matches anchor 3 ("some concrete guidance but incomplete") rather than anchor 4, which expects concrete code or commands with only minor gaps.

3 / 5

Workflow Clarity

快速指导 gives a coherent numbered routing sequence (understand focus → responsibility split → route by problem type → handoff to sibling skills), which is clear for this non-destructive navigation task. It is below anchor 5 because there are no checkpoints for ambiguous cases (e.g., a problem spanning both Agent and Worker) and no guidance on what to do after opening a reference doc.

4 / 5

Progressive Disclosure

The body is a well-structured overview with clearly signaled, one-level-deep references organized by problem type, matching anchor 4. It does not reach anchor 5 because the cited paths (reference/1-runtime-upgrade.md, reference/2-ask-heartbeat-build.md, reference/3-platform-ops.md) do not exist in the evaluated bundle, and the reference list is duplicated between 快速指导 and 延伸阅读.

4 / 5

Total

15

/

20

Passed

Description

87%Weight 40%Scale 1-5

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

A strong description: it states a concrete capability inventory, gives an explicit and disambiguating use-when clause, and uses natural domain terminology. The only weaknesses are minor — a few missing capability mentions and synonym coverage in trigger terms.

DimensionReasoningScore

Specificity

The description enumerates several concrete capability areas ("守护进程、心跳、Ask 轮询、任务拉起、升级更新和与 Dispatch/Worker 的协作"), giving specific scope with only minor gaps (e.g., environment/ops topics are not mentioned). It names capability areas rather than fully comprehensive concrete actions, so it sits at anchor 4 rather than 5.

4 / 5

Completeness

It explicitly answers both "what" (the enumerated Agent build-machine-side capabilities) and "when" ("当用户要改构建机宿主侧行为而不是 Worker 执行细节时优先使用") with concrete trigger phrasing, matching anchor 5. It is not anchor 4 because the "when" clause is fully explicit and includes disambiguating conditions rather than being merely adequate.

5 / 5

Trigger Term Quality

Terms like "守护进程", "心跳", "Ask 轮询", "任务拉起", "升级更新", and "Dispatch/Worker" are exactly what a BK-CI developer would say when needing this skill. Coverage is good but misses some natural synonyms and variations (e.g., online-rate, build-machine environment, English equivalents), matching anchor 4 rather than the comprehensive anchor 5.

4 / 5

Distinctiveness Conflict Risk

The description carves out a clear niche (BK-CI Agent build-machine host side) and explicitly disambiguates against the closest sibling ("而不是 Worker 执行细节"), minimizing conflict risk. This matches anchor 5's "clear niche with distinct triggers; minimal conflict risk".

5 / 5

Total

18

/

20

Passed

Validation

100%

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
TencentBlueKing/bk-ci
Reviewed

Table of Contents

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.