darwinian-evolver

Evolve prompts/regex/SQL/code with Imbue's evolution loop.

Quality

71%

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 ./optional-skills/research/darwinian-evolver/SKILL.md

Quality

Content

92%

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 strong, lean, executable skill: concise, fully actionable with real shipped scripts, and a well-sequenced workflow with an explicit verification checkpoint. The only notable weakness is progressive disclosure — a referenced templates/ file does not exist in the bundle, leaving one broken path.

Suggestions

Add the referenced templates/custom_problem_template.py to the bundle, or remove the 'The skill ships templates/custom_problem_template.py' line to avoid a broken reference.

Consider moving the hyperparameter table or pitfalls into a reference file so SKILL.md stays an overview, since the rubric favors one-level-deep splitting for non-trivial skills.

Verify the OpenRouter quick-start command's SKILL_DIR path matches the actual install location so the copy-paste example resolves.

Dimension	Reasoning	Score
Conciseness	The body is lean and assumes Claude's competence: no preamble explaining what evolution or LLMs are, every section earns its place (install, quick start, hyperparameter table, pitfalls), and it avoids the padded-concept style of the level-1 anchor.	3 / 3
Actionability	Copy-paste-ready bash and Python commands throughout (install, parrot smoke test, OpenRouter driver invocation, snapshot inspection), concrete flag tables with defaults and change-guidance, and a fully shipped reference driver in scripts/parrot_openrouter.py — fully executable rather than pseudocode.	3 / 3
Workflow Clarity	Install → verify → smoke-test → custom-problem is clearly sequenced, with an explicit 'Verification' checkpoint ('exit code 0 from this is sufficient') and a numbered three-part Problem definition; the body also warns about destructive/batch pitfalls (aggressive concurrency, AGPL import) which is the feedback-loop care the rubric rewards.	3 / 3
Progressive Disclosure	Structure is good and references to real bundle files are clearly signaled (scripts/parrot_openrouter.py, scripts/show_snapshot.py both exist), but the body also references 'templates/custom_problem_template.py' which is not present in the bundle, and there are no references/ or assets/ — so a referenced path is broken, capping this at 2 rather than a clean level-3 split.	2 / 3
	Total	11 / 12 Passed

Description

50%

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 is concise and third-person with a concrete domain, but it omits any explicit 'when to use' trigger clause and natural trigger phrases that the body actually contains, capping completeness and trigger-term quality at 2. It is a clear level-2 description: accurate and specific enough, but not triggering guidance-rich.

Suggestions

Add an explicit trigger clause, e.g. 'Use when the user asks to optimize a prompt, evolve a regex/SQL query, or auto-improve a small code snippet against a scorer.'

Surface natural trigger terms the user would actually say ('optimize this prompt', 'auto-improve this code', 'evolve a regex') into the description rather than only listing artifact types.

Distinguish the skill from generic prompt-engineering by naming the fitness-function prerequisite in the description.

Dimension	Reasoning	Score
Specificity	Names a concrete domain ('prompts/regex/SQL/code') and the action ('Evolve') but does not enumerate the multiple specific operations (optimize, search, auto-improve) found in the body, so it sits at 'names domain and some actions, but not comprehensive' rather than the multi-action level 3.	2 / 3
Completeness	It clearly answers 'what does this do' ('Evolve prompts/regex/SQL/code with Imbue's evolution loop') but there is no 'when should Claude use it' clause — the rubric caps completeness at 2 when an explicit 'Use when...' trigger is missing.	2 / 3
Trigger Term Quality	'prompts/regex/SQL/code' are relevant keywords a user might mention, but there are no natural trigger phrases users would actually say ('optimize this prompt', 'evolve a regex'), missing common variations that the body itself contains.	2 / 3
Distinctiveness Conflict Risk	'Imbue's evolution loop' carves a fairly distinct niche unlikely to conflict broadly, but without explicit trigger guidance it could still overlap with general prompt-engineering or optimization skills, so it is 'somewhat specific but could still overlap' rather than a clear level-3 niche.	2 / 3
	Total	8 / 12 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: NousResearch/hermes-agent
Path: optional-skills/research/darwinian-evolver/SKILL.md
Commit: 4c9628e

Reviewed: about 14 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.