Name: breezing
Rating: 48 (1 reviews)
Author: Chachamaru127

breezing

Team execution mode — backward-compatible alias for harness-work with team orchestration. Composer/composer 2.5 maps to the cursor backend.

Quality

51%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./opencode/skills/breezing/SKILL.md

The canonical home for this skill is breezing in Chachamaru127/claude-code-harness

Quality

Content

66%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 content is a thorough, well-sequenced operational protocol with strong workflow clarity and concrete commands, but it is verbose for an alias skill — inline version/date noise and restated cursor-path detail cost token efficiency.

Suggestions

Move inline version-specific notes (CC 2.1.x, Codex 0.123.0, dated operator rulings) into a clearly labeled 'Version notes / deprecated' section or a reference file so they stop penalizing the main flow.

Consolidate the cursor / lean path, which is currently restated in Quick Reference, the Options table, and the Cursor Backend Fast Path section, into a single authoritative location.

Consider extracting the detailed reviewer-only mode and codex mode walkthroughs into a reference file, keeping SKILL.md a leaner overview that points to them.

Dimension	Reasoning	Score
Conciseness	The body is noticeably verbose for an alias skill: inline time-sensitive version/date notes (CC 2.1.111/2.1.108+/2.1.98+/2.1.217/2.1.219, Codex 0.123.0, operator 裁定 2026-07-24) sit outside any deprecated section, and the cursor/lean path is restated across the Quick Reference, Options table, and a dedicated section.	2 / 5
Actionability	It gives mostly executable guidance — concrete commands like `codex-companion.sh task --write`, `bin/harness work-mode on`, `resolve-impl-backend.sh`, plus banner/JSON templates — with only minor gaps (placeholders like <wt>, <base_ref> and a few pseudocode flow blocks).	4 / 5
Workflow Clarity	The Plan→Work→Integrated Review Gate→Finalize pipeline is explicitly sequenced with validation checkpoints (self_review gate, brain verdict, 'APPROVE が出るまで反復、最大3回', human escalation) and feedback loops for a batch/destructive operation, satisfying the rubric's feedback-loop requirement.	5 / 5
Progressive Disclosure	Structure is good: the two real reference files (lean-path-detail.md, monitor-and-learning.md) are well-signaled one level deep via explicit links, but the ~450-line SKILL.md keeps some detailed flows (reviewer-only mode, codex mode) inline that could live in references.	4 / 5
	Total	15 / 20 Passed

Description

37%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 identifies the skill's niche but is jargon-heavy, lists no concrete actions, and omits any 'Use when...' trigger guidance, which caps completeness. It reads as an internal wiring note rather than a discoverable, user-facing trigger.

Suggestions

Add an explicit 'Use when...' clause with natural trigger phrases (e.g., 'Use when running multiple harness tasks end-to-end as a coordinated team, or when the user asks to delegate implementation to Workers with an independent Reviewer').

Replace jargon-only terms with at least one or two natural keywords a user would say ('run all tasks', 'delegate to workers', 'team review loop').

State 1-2 concrete actions the skill performs (plans, delegates to Workers, runs an independent Reviewer, cherry-picks, reports) instead of only describing it as an alias.

Dimension	Reasoning	Score
Specificity	It names the domain ('Team execution mode', 'alias for harness-work with team orchestration') but lists no concrete actions a user can recognize — 'team orchestration' is a generic capability descriptor, and 'Composer/composer 2.5 maps to the cursor backend' is a mapping fact rather than an action.	2 / 5
Completeness	There is a moderately clear 'what' (a team execution mode aliasing harness-work) but no 'when' — the missing 'Use when...' trigger guidance caps completeness at 3 per the rubric.	3 / 5
Trigger Term Quality	The terms present ('harness-work', 'Composer/composer 2.5', 'cursor backend') are internal harness jargon; there are no natural phrases a user would actually say to invoke this, landing noticeably below the midpoint between anchor 1 and anchor 3.	2 / 5
Distinctiveness Conflict Risk	It occupies a clear niche (harness team orchestration / cursor-composer backend) but, being an explicit 'backward-compatible alias for harness-work', it carries real overlap risk with that closely related skill rather than only minor risk.	3 / 5
	Total	10 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
relative_links	Relative link issues: 7 missing, 7 deeper-than-1-level, 1 suspicious	Warning
referenced_paths_exist	Referenced path issues: 4 missing	Warning

	Total	14 / 16 Passed

Repository: Chachamaru127/claude-code-harness
Path: opencode/skills/breezing/SKILL.md
Commit: 7d8deba

Reviewed: about 15 hours 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.