aris-experiment-bridge

Workflow 1.5: Bridge between idea discovery and auto review. Reads EXPERIMENT_PLAN.md, implements experiment code, deploys to GPU, collects initial results. Use when user says "实现实验", "implement experiments", "bridge", "从计划到跑实验", "deploy the plan", or has an experiment plan ready to execute.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

High

Do not use without reviewing

Quality

Content

77%

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

A highly actionable, clearly sequenced multi-phase workflow with strong validation checkpoints. Its main weakness is token efficiency: prose warning blocks and inline templates that could be trimmed or split into referenced files inflate the body.

Suggestions

Tighten or remove prose warning/dialogue blocks and rationale sentences (e.g. the COMPUTE UNAVAILABLE banner, 'Writing code without a deployment target wastes time...') to reduce padding and improve conciseness.

Move the large reusable templates — the GPT-5.4 code-review MCP prompt (Phase 2.5) and the Initial Experiment Results markdown (Phase 5) — into reference files in references/ and link to them one level deep, improving progressive_disclosure.

Condense the Key Rules list by merging overlapping lifecycle/budget rules so every remaining line earns its place in the context window.

Dimension	Reasoning	Score
Conciseness	The ~340-line body is mostly efficient and assumes Claude's competence, but contains padded prose — warning dialogue blocks and rationale sentences like 'Writing code without a deployment target wastes time and may lead to hallucinated results' — that could be tightened, so it does not reach the lean/efficient anchor.	2 / 3
Actionability	Provides fully executable, specific commands throughout — 'nvidia-smi --query-gpu=index,memory.used,memory.total --format=csv,noheader', 'git clone <BASE_REPO>', 'python train.py --config configs/run_id.yaml --seed 42', and concrete MCP configs — copy-paste ready.	3 / 3
Workflow Clarity	A clear Phase 0→6 sequence with explicit validation checkpoints (compute guard halt, sanity pass criteria, review re-submit loop capped at 2 rounds, auto-debug max 3 attempts) and feedback loops for fragile operations.	3 / 3
Progressive Disclosure	No bundle files exist and the skill is a monolithic single file; large inline templates (the code-review MCP prompt, the results-summary markdown) that could be split into referenced files remain inline, fitting the 'content that should be separate is inline' anchor.	2 / 3
	Total	10 / 12 Passed

Description

100%

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, specific description that covers concrete actions, natural bilingual trigger terms, both the 'what' and 'when', and a distinct niche. It does not over-claim or pad with fluff.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Reads EXPERIMENT_PLAN.md, implements experiment code, deploys to GPU, collects initial results' — matching the anchor for several specific concrete actions rather than only naming a domain.	3 / 3
Completeness	Explicitly answers both what it does ('Reads...implements...deploys...collects') and when to use it via a clear 'Use when...' clause with concrete triggers, matching the top anchor.	3 / 3
Trigger Term Quality	Provides good coverage of natural phrases a user would say, bilingual — '实现实验', 'implement experiments', 'bridge', '从计划到跑实验', 'deploy the plan', plus 'experiment plan ready to execute'; 'bridge' alone is a touch generic but overall coverage is strong.	3 / 3
Distinctiveness Conflict Risk	Occupies a clear niche as the Workflow 1.5 experiment bridge with distinct bilingual triggers, making it unlikely to fire for the wrong skill.	3 / 3
	Total	12 / 12 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
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: OpenLAIR/dr-claw
Path: skills/aris-experiment-bridge/SKILL.md
Commit: 6da7f7c

Reviewed: about 8 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.