devtu-self-evolve

Orchestrate the full ToolUniverse self-improvement cycle: discover APIs, create tools, test with researcher personas, fix issues, optimize skills, and push via git. References and dispatches to all other devtu skills. Use when asked to: run the self-improvement loop, do a debug/test round, expand tool coverage, improve tool quality, or evolve ToolUniverse.

Quality

88%

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

Quality

Content

85%

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 orchestrator body with strong validation checkpoints and clean one-level-deep references. The main weakness is conciseness — some inline rubric/category content could move to reference files to respect the stated line budget.

Suggestions

Move the 'Skill Usefulness Testing' rubric and 'Common Issue Categories' table into a reference file (e.g. references/usefulness-testing.md) and link to it, to bring the body back under its 150-line target.

The 'Common failure patterns' table largely duplicates references/bug-patterns.md — replace it with a one-line pointer to keep the overview lean.

Dimension	Reasoning	Score
Conciseness	The body is mostly lean and operational, but the embedded usefulness-testing rubric table and the common-issue categories partly duplicate references/bug-patterns.md, and the file exceeds its own stated 150-line budget.	2 / 3
Actionability	It provides copy-paste-ready commands throughout — 'gh pr list --state open', 'python3 -m tooluniverse.cli run <ToolName> \'<json_args>\'', 'ruff check src/tooluniverse/<file>.py', and full benchmark script paths — fully executable guidance.	3 / 3
Workflow Clarity	The six phases are clearly sequenced with explicit validation checkpoints (CRITICAL CLI verification, Phase 4 lint/test/verify, Phase 6 mergeable check) and feedback loops (pre-commit stage→commit→re-stage→commit).	3 / 3
Progressive Disclosure	The body is an overview with well-signaled one-level-deep references to real bundle files ([references/persona-template.md], [references/bug-patterns.md]), both of which exist, keeping detail appropriately split out.	3 / 3
	Total	11 / 12 Passed

Description

92%

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, third-person description with explicit concrete actions and a clear 'Use when' trigger clause. The only real weakness is overlap between its trigger terms and the dispatched devtu sub-skills it routes to.

Suggestions

Narrow the 'Use when' triggers to orchestrator-level intents (e.g. 'run the full self-improvement cycle', 'evolve ToolUniverse') and let sub-skills own sub-intents like 'fix a broken tool' to reduce sibling-skill conflicts.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'discover APIs, create tools, test with researcher personas, fix issues, optimize skills, and push via git' — matching the anchor for several specific actions rather than vague language.	3 / 3
Completeness	Explicitly answers both 'what' (orchestrate the cycle via the listed actions) and 'when' (an explicit 'Use when asked to...' trigger clause), so it is not capped at 2.	3 / 3
Trigger Term Quality	The 'Use when asked to: run the self-improvement loop, do a debug/test round, expand tool coverage, improve tool quality, or evolve ToolUniverse' clause gives good coverage of natural phrases a user would say.	3 / 3
Distinctiveness Conflict Risk	It has a clear orchestrator niche, but several triggers ('expand tool coverage', 'improve tool quality', 'fix a broken tool') overlap the devtu sub-skills it dispatches to, so it could still trigger for a sibling skill.	2 / 3
	Total	11 / 12 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: mims-harvard/ToolUniverse
Path: skills/devtu-self-evolve/SKILL.md
Commit: 089eb8e

Reviewed: about 7 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.