receive-handoff-grounding

接球前真相核验三问：claim → resolver → verdict (sourceTier T0/T1/T2 + actionFamily)，防止把传球者当无审视真相源（F167 Phase O 第一性原理）。 Use when: 即将调 hold_ball / register_pr_tracking / register_issue_tracking / merge / takeover / 改 owner / 任何 irreversible action / 基于 "operator signoff" 或 "你是 owner" 类 claim 行动之前。 Not for: 纯阅读 cross_post（无 actionFamily 后续）；本 thread 日常 @mention 无副作用； implementation continuation（自检通过的下一步）。 Output: claim grounding verdict (verified/mismatch/insufficient) + 接球决策 (proceed / block / push back to source thread)。

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Medium

Suggest reviewing before use

Quality

Content

77%Weight 40%Scale 1-3

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

A highly actionable, well-sequenced reflex workflow with concrete schemas, resolver calls, and validation checkpoints. Its weaknesses are length/redundancy and progressive disclosure: critical detail is pointed at refs/*.md files that are not bundled.

Suggestions

Create the referenced refs/resolver-catalog.md, refs/claim-schema.md, and refs/dogfood-fixtures.md (or remove the references and inline only what the overview needs) so progressive-disclosure navigation resolves.

De-duplicate the trigger list, which currently appears in the frontmatter, the hard-trigger table, the self-activation detection table, and the checklist — keep one canonical source.

Move the 8 dogfood demos to refs/dogfood-fixtures.md and retain only 1–2 representative examples inline to cut length without losing the failure-mode coverage.

Dimension	Reasoning	Score
Conciseness	Content is information-dense rather than padded with concepts Claude knows, but at ~300 lines it restates rules several times (triggers reappear in the body table, self-activation table, and checklist) and the 8 demos re-explain already-stated policies, so it could be tightened.	2 / 3
Actionability	Gives executable guidance: a TypeScript WaitSourceRef schema, concrete resolver calls ('cat_cafe_get_message(messageId).author === \"you\"', 'feat_index.lookup(...)', 'git log --grep --author', 'gh api'), an action policy table, and a copy-paste push-back template.	3 / 3
Workflow Clarity	The Q1/Q2/Q3 reflex is a clear sequenced workflow with explicit validation checkpoints (the '自检 checklist 每次接球必过', the ≥1 T0/T1 evidence rule) and feedback loops (mismatch → block + push back; insufficient → fail-closed), covering the destructive/irreversible actions it governs.	3 / 3
Progressive Disclosure	Sections are clearly labeled and references are signaled ('refs/resolver-catalog.md', 'refs/claim-schema.md', 'refs/dogfood-fixtures.md'), but those referenced files do not exist in the bundle and substantial detail (schema, 8 demos, telemetry) lives inline, so navigation to the offloaded material is broken.	2 / 3
	Total	10 / 12 Passed

Description

85%Weight 40%Scale 1-3

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 dense, well-structured description that clearly states what, when, not-for, and output with concrete action tokens. Its only weakness is trigger phrasing built from internal tool names rather than natural user speech.

Suggestions

Add a few natural-language trigger variants a user might actually say (e.g. 'merge this PR', 'take over this task', 'wait for the reporter') alongside the internal tool names to raise trigger_term_quality.

Consider trimming the F167/sourceTier jargon from the lead sentence so the first line reads as a capability statement before the mechanism.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions and objects — 'claim → resolver → verdict (sourceTier T0/T1/T2 + actionFamily)', 'hold_ball / register_pr_tracking / register_issue_tracking / merge / takeover / 改 owner / 任何 irreversible action' — rather than vague language.	3 / 3
Completeness	Explicitly answers what (核验三问 claim→resolver→verdict + 接球决策), when ('Use when: 即将调 ...'), what-not-for ('Not for:'), and output ('Output: claim grounding verdict ...'), so both what and when are clearly stated.	3 / 3
Trigger Term Quality	Coverage of the domain's trigger tokens is broad (hold_ball, register_*_tracking, merge, takeover, 改 owner, irreversible, operator signoff), but these are internal tool/actionFamily jargon, not phrases a user would naturally say; common natural variations are missing.	2 / 3
Distinctiveness Conflict Risk	Occupies a clear niche — pre-handoff claim grounding for a2a cat-cafe handoffs — with highly specific triggers (register_*_tracking, hold_ball, takeover, irreversible) unlikely to fire for unrelated 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: zts212653/clowder-ai
Path: cat-cafe-skills/receive-handoff-grounding/SKILL.md
Commit: 80782c5

Reviewed: 11 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.