Tessl • The management layer for AI agent skills

The body is an exceptionally dense, purely operational orchestration guide: exact commands, explicit state-field checks, conditional fetch rules to avoid redundant state reads, and thorough pass/fail/unknown/recover handling with feedback loops. Its only structural weakness is that every referenced detail file (including the locally-referenced compat.md) is absent from the shipped bundle, so progressive disclosure is well designed on paper but unverifiable in practice.

Dimension	Reasoning	Score
Conciseness	The body contains no concept explanations or filler — every sentence is operational, e.g., "确认 `data.runtimeFormat: sdk`、`data.phase: verify` 和 `nextAction.kind`" — and reload guards like "本会话已有且仍有效时直接复用" actively prevent wasted work. Nothing trimmable stands out.	5 / 5
Actionability	Six exact CLI invocations with flags and arguments are given ("comet state transition <change-name> verify-fail --reason", "comet guard <change-name> verify --report <report-ref> --apply -- <program> [args...]") alongside explicit JSON-field checks and per-case routing for pass, fail, and recover; placeholders are standard parameterization rather than gaps.	5 / 5
Workflow Clarity	Numbered steps 1-3 are gated by explicit validation checkpoints ("验证结论合格后", "Guard 通过后", "检查失败后读取同次响应的失败收据") with a real feedback loop (verify-fail transition returns to Build and continues from the current Build Action), and ambiguous states are handled deliberately ("结果未明时停止").	5 / 5
Progressive Disclosure	The file is a lean overview with well-signaled, one-level-deep references to sibling-skill docs (sdk-methods.md, edit-review.md, change-walkthrough.md) and "reference/compat.md" for the legacy path. However, the bundle ships no reference files at all, so the local compat.md link and the cross-skill targets cannot be verified to resolve, leaving minor navigation gaps.	4 / 5
	Total	19 / 20 Passed

Description

73%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.

The description is a compact, third-person statement that answers both 'what' (verify a Classic change and record results) and 'when' (explicit slash-command and runtime-phase triggers) with low conflict risk. Its main weakness is shallow capability coverage — the many concrete verification activities in the body are compressed into two high-level verbs. Adding natural-language trigger variations would round out the keyword coverage.

Suggestions

Enumerate the phase's concrete capabilities in the description (e.g., risk assessment, final integration review, running light/full verification, writing the verification report, submitting guard receipts) to raise specificity beyond 'verify and record'.

Add natural-language trigger phrases a user might say besides the slash command (e.g., "验证变更结果", "verify the change") so it triggers on plain requests, not only /comet-verify or runtime routing.

State in the what-clause that results are recorded as a guard-submitted verification report, which sharpens both the 'what' and the skill's distinctiveness from generic verification helpers.

Dimension	Reasoning	Score
Specificity	"验证 Classic change 并记录结果" names the domain and exactly two concrete actions (verify, record results) but omits the phase's other capabilities (risk assessment, final integration review, report submission via guard, optional walkthrough), matching the '1-2 concrete actions, not comprehensive' anchor rather than the several-action anchors above.	3 / 5
Completeness	Both parts are present: a clear what ("验证 Classic change 并记录结果") and an explicit when clause with concrete triggers ("在用户调用 /comet-verify...时使用"). It sits below the top anchor only because the 'what' is thin — two high-level actions versus the multi-action anchor-5 example.	4 / 5
Trigger Term Quality	The triggers "用户调用 /comet-verify，或 Classic Runtime 进入 Verify" cover the exact slash command a user would type plus the runtime phase trigger, but natural plain-language variations (e.g., "验证这个变更" or "check the change") are missing, keeping it below the comprehensive-synonyms anchor.	4 / 5
Distinctiveness Conflict Risk	"Classic change", "/comet-verify", and "Classic Runtime 进入 Verify" carve out a clear niche (the Verify phase of the Comet Classic pipeline) with an exact, unique command trigger, so overlap with unrelated or sibling skills is minimal — the phase name and command disambiguate cleanly.	5 / 5
	Total	16 / 20 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
relative_links	Relative link issues: 1 missing, 4 suspicious	Warning

	Total	15 / 16 Passed

Repository: rpamis/comet
Path: assets/skills-zh/comet-verify/SKILL.md
Revision: a61a2e5

Reviewed: 5 days ago

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.

comet-verify

Table of Contents

Is this your skill?