inno-experiment-dev

Creates implementation plan, writes project code with judge feedback loop, and submits final experiment run. Use after code-survey in both Idea and Plan branches.

Quality

67%

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 ./skills/inno-experiment-dev/SKILL.md

Quality

Content

75%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 well-structured, information-dense orchestration spec with a clear phased workflow, a judge feedback loop, concrete function calls, and appropriately signaled references to real bundle files. Main weaknesses are repeated path-resolution disclaimers, missing prompts/*.md files, and a weaker validation checkpoint in the submit phase.

Suggestions

Consolidate the repeated Dr. Claw path-resolution note into a single 'Path resolution' subsection instead of restating it in the inputs table and steps 5 and 7a.

Either add the missing prompts/*.md files to the bundle (prompts/build_plan_query.md, build_ml_dev_query.md, etc.) or note that they live elsewhere so references are not dangling.

Add an explicit validation checkpoint before the final submit run (e.g. verify checkpoints/tests pass before calling the ML Agent with iter_times='submit') to strengthen the submit-phase feedback loop.

Dimension	Reasoning	Score
Conciseness	The body is dense and assumes Claude's competence (function names, line-number citations, message structures), but repeats the path-resolution disclaimer ('absolute in Dr. Claw–created projects; use as-is or resolve if relative') across the inputs table and steps 5 and 7a, which could be trimmed.	4 / 5
Actionability	Provides concrete function signatures (build_plan_query, build_ml_dev_query, etc.), exact message formats, and a tool-mapping table, but the referenced prompts/*.md files are absent from the bundle, leaving minor gaps in executable detail.	4 / 5
Workflow Clarity	A clearly sequenced 3-phase/11-step workflow with an explicit judge-feedback loop (iterate until 'fully_correct': true) and a closing checklist; the submit phase relies on case_resolved/case_not_resolved without an explicit pre-submit validation checkpoint, a minor gap.	4 / 5
Progressive Disclosure	Good structure with one-level-deep, signaled references to real bundle files (references/coding_plan_agent.md, ml_agent_instructions.md, judge_agent_instructions.md), though referenced prompts/*.md files do not exist in the bundle, a minor organization gap.	4 / 5
	Total	16 / 20 Passed

Description

60%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 clearly states what the skill does across three concrete phases and when to use it within the pipeline, but its trigger language is internal pipeline jargon rather than natural user terms. Specificity and completeness are solid; trigger-term quality is the main weakness.

Suggestions

Add natural-language trigger phrases users would actually say (e.g. 'experiment implementation', 'training run', 'ML dev iteration') rather than only pipeline-internal cues like 'Use after code-survey'.

Reframe the 'when' clause to include user-facing triggers in addition to pipeline ordering, e.g. 'Use when implementing and iterating on an experiment plan after code-survey'.

Consider mentioning concrete artifacts users care about (model checkpoints, statistical outputs, submission results) to sharpen specificity toward a 5.

Dimension	Reasoning	Score
Specificity	Lists three concrete actions ('Creates implementation plan, writes project code with judge feedback loop, and submits final experiment run'), giving clear coverage of the skill's phases with only minor gaps versus a fully comprehensive list.	4 / 5
Completeness	It answers 'what' with three concrete actions and gives an explicit 'when' ('Use after code-survey in both Idea and Plan branches'), but the 'when' is pipeline-positioning rather than user-trigger language and could be more explicit.	4 / 5
Trigger Term Quality	The only trigger cues are pipeline-internal jargon ('Use after code-survey in both Idea and Plan branches') rather than natural phrases a user would say, so it barely registers natural keywords.	2 / 5
Distinctiveness Conflict Risk	The experiment-dev niche and 'Use after code-survey' positioning are mostly distinct, though sibling inno pipeline skills (idea-generation, code-survey, experiment-analysis) create minor overlap risk.	4 / 5
	Total	14 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: OpenLAIR/dr-claw
Path: skills/inno-experiment-dev/SKILL.md
Commit: 4b35756

Reviewed: 5 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.