Aeon Skill Evals
BankrBot/skills
Validate the output of any installed skill against an assertion manifest — word counts, required patterns, forbidden phrases, required sections, source citation.
Improves Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with held-out gates.
Install Skill Eval Improve
$ npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install Arenukvern/mcp_flutter skill-eval-improve --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/Arenukvern/mcp_flutter.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/skill-eval-improve .claude/skills/skill-eval-improve && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "skill-eval-improve" agent skill from https://github.com/Arenukvern/mcp_flutter/tree/main/.agents/skills/skill-eval-improve into .claude/skills/skill-eval-improve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "skill-eval-improve", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/Arenukvern/mcp_flutter/tree/main/.agents/skills/skill-eval-improveType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install Arenukvern/mcp_flutter skill-eval-improve --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Arenukvern/mcp_flutter.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/skill-eval-improve .agents/skills/skill-eval-improve && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "skill-eval-improve" agent skill from https://github.com/Arenukvern/mcp_flutter/tree/main/.agents/skills/skill-eval-improve into .agents/skills/skill-eval-improve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "skill-eval-improve", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install Arenukvern/mcp_flutter skill-eval-improve --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Arenukvern/mcp_flutter.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/skill-eval-improve .cursor/skills/skill-eval-improve && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "skill-eval-improve" agent skill from https://github.com/Arenukvern/mcp_flutter/tree/main/.agents/skills/skill-eval-improve into .cursor/skills/skill-eval-improve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "skill-eval-improve", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/Arenukvern/mcp_flutter.git --path .agents/skills/skill-eval-improve--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install Arenukvern/mcp_flutter skill-eval-improve --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Arenukvern/mcp_flutter.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/skill-eval-improve .gemini/skills/skill-eval-improve && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "skill-eval-improve" agent skill from https://github.com/Arenukvern/mcp_flutter/tree/main/.agents/skills/skill-eval-improve into .gemini/skills/skill-eval-improve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "skill-eval-improve", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install Arenukvern/mcp_flutter skill-eval-improveInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/Arenukvern/mcp_flutter.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/skill-eval-improve .github/skills/skill-eval-improve && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "skill-eval-improve" agent skill from https://github.com/Arenukvern/mcp_flutter/tree/main/.agents/skills/skill-eval-improve into .github/skills/skill-eval-improve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "skill-eval-improve", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install Arenukvern/mcp_flutter skill-eval-improve --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Arenukvern/mcp_flutter.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/skill-eval-improve .opencode/skills/skill-eval-improve && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "skill-eval-improve" agent skill from https://github.com/Arenukvern/mcp_flutter/tree/main/.agents/skills/skill-eval-improve into .opencode/skills/skill-eval-improve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "skill-eval-improve", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
skill-eval-improve- GitHub stars
- 386
- Token cost
- ~2.4k tokens
- SKILL.md length
- 827 words
- Files
- 12 (incl. references)
- Skills in repo
- 23
- Repo updated
- First seen
- Licence
- MIT
At a glance
Improves Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with held-out gates.
- Works in 4 steps: Write 3–5 representative user prompts… → Run agent without skill → record failures. → Run with skill → record improvements and… → …
- Tuning skill quality
- SKILL.md covers When to use, When not to use, Mixture of experts (evaluation… and Layer 0 — Skill Steward…, plus 13 more sections
- Calls pnpm and npx
What it does
Skill Eval Improve is an agent skill from Arenukvern/mcp_flutter. Improves Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with held-out gates. Use when tuning skill quality, routing, or adopting Chrome/Microsoft T-named quality gates—not for bulk validate-only or SkillOpt automation.
Its SKILL.md is about 2.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 14 other files, including reference files (for example `evals/cases/duplicated-guidance-compression-trigger.yaml`, `evals/cases/improve-routing-trigger.yaml` and `evals/cases/tool-loop-drift-trigger.yaml`).
It sits in Agent Workflows, covering Agent evaluation and testing, Quality gates and LLM evaluation. It works with pnpm. The repository describes itself as: MCP Toolkit for Flutter AI Agent Driven Development (MCP/CLI + custom client side tools) - via closed feedback loop (visual & semantic snapshot) and high client side… The licence is MIT.
When your agent uses it
- Tuning skill quality
- Adopting Chrome/Microsoft T-named quality gates—not for bulk validate-only
- SkillOpt automation
Example prompts
- “Use the skill-eval-improve skill to improve Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with…”
- “/skill-eval-improve”
Requirements
- Node.js
Workflow steps
4 steps, taken from the first numbered list in SKILL.md.
- Write 3–5 representative user prompts (should trigger + should not trigger).
- Run agent without skill → record failures.
- Run with skill → record improvements and new failures.
- Mirror prompts in evals/cases/*.yaml (CI rules) and references/evals.md (behavior log).
What it can do on your machine
Read from SKILL.md and the folder at commit 62f3ee1. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
pnpmnpx
From the folder's file list and the shell code blocks in SKILL.md.
Network
Links to these hosts (documentation or services it may open):
microsoft.github.ioarxiv.orggithub.com
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Skill Eval Improve loads about 2.4k tokens when it runs, and up to ~5.4k if it reads all its reference files. Until then it costs about 71 tokens; SKILL.md has 827 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from Arenukvern/mcp_flutter at commit 62f3ee1, republished under its MIT licence (© Arenukvern). 827 words, ~2,388 tokens.
Download SKILL.mdSave it as
.claude/skills/skill-eval-improve/SKILL.md (or your agent's skills folder). This skill also uses 11 other files; get the full folder from GitHub.- name
- skill-eval-improve
- description
- Improves Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with held-out gates. Use when tuning skill quality, routing, or adopting Chrome/Microsoft T-named quality gates—not for bulk validate-only or SkillOpt automation.
- license
- MIT
- type
- governance
- metadata.author
- skill-steward
- metadata.version
- 1.1.0
- metadata.category
- marketplace
- paths
- skills/**/evals/**, scripts/eval-skill.mjs, scripts/eval-tiers.mjs
Skill eval & improve
Improve skills measurably: baseline → measure → bounded edit → re-validate. Combine local tooling, Codex plugin-eval (when installed), and research-backed loops (SkillOpt).
When to use
- Skill triggers wrong or never loads (description routing)
- Bloated
SKILL.md, high token cost, weak outcomes - After adding a new procedure—need regression checks
- Porting patterns from product MCP / plugin-eval research into Skill Steward skills
When not to use
- Bulk repo validation — e.g. “validate every skill in this repo” →
pnpm run validateonly (skill-authoring-lifecycle for audit); do not start benchmark or SkillOpt loops. - Automated SkillOpt / cluster training — Skill Steward documents a manual bounded-edit loop; no overnight optimizer pipeline.
- Creating a new skill — use skill-authoring-lifecycle first; eval-improve applies after a skill exists.
Cursor scope (optional): activate when editing under skills/** or scripts/validate-skills.mjs.
Mixture of experts (evaluation stack)
| Layer | Expert | Tool / method | Cost |
|---|---|---|---|
| 0 — Gate | Lint | pnpm run validate, skill-authoring-lifecycle | seconds |
| 0b — Rules | Routing/docs SSOT | pnpm run eval (T1 behavior-critical YAML cases) | seconds |
| 1 — Static | Structure | Codex plugin-eval analyze (if available) | seconds |
| 2 — Human | Behavior | 3–5 prompts with/without skill | minutes |
| 3 — Measured | Usage | plugin-eval benchmark + measurement-plan | minutes–hours |
| 4 — Evolve | Text optimization | SkillOpt-style bounded edits + held-out gate | hours |
| 5 — Navigate | Telemetry / dogfood | Current steward benchmark scenarios on compact traces | seconds |
Use the cheapest layer that answers the question. Do not skip layer 0.
Layer 0 — Skill Steward validator (always)
pnpm run validate
pnpm run validate:json # CI / automationFix all error: lines. Treat warn: (missing sources.md, long SKILL.md) seriously.
T-named skill quality gates (ADR 0011, ADR 0027)
| Tier | Skills | CI |
|---|---|---|
| T1 — Behavior-critical eval-gated | Routing/procedure skills where drift can change agent decisions, claims, delegation, governance, or evidence boundaries | pnpm run eval + validate |
| T2 — Structural validate-only | All others | pnpm run validate |
T1 behavior-critical currently includes harness-engineering-lifecycle, mcp-harness-repo-maintainer, mixture-of-experts, multi-agent-handoff, plugin-marketplace-setup, repo-quality-system-lifecycle, repository-governance-lifecycle, skill-authoring-lifecycle, skill-eval-improve, steward-continuity-boundary-lifecycle, and vision-alignment-foresight. Each requires evals/cases/*.yaml (≥2) + references/evals.md. Schema: eval-case-schema.md.
Layer 0b — Rule-based cases (T1 behavior-critical CI)
pnpm run eval
pnpm run eval -- --skill mcp-harness-repo-maintainer
pnpm run eval:jsonChrome eval design (failure modes, rubrics, objective vs judge): references/chrome-eval-design.md.
CI does not run LLM judges. Subjective quality stays in references/evals.md (layer 2+).
Layer 1 — Codex plugin-eval (local)
When Codex plugin-eval is installed (~/.codex/plugins/.../plugin-eval):
# Chat-first router
plugin-eval start skills/<name> --request "Evaluate this skill." --format markdown
# Static report
plugin-eval analyze skills/<name> --format markdown
# Token budget explanation
plugin-eval explain-budget skills/<name> --format markdown
# Starter benchmark config
plugin-eval init-benchmark skills/<name>
plugin-eval benchmark skills/<name> --dry-runHand off rewrite plans to plugin-eval’s improve-skill skill after analyze --brief-out.
Details: references/plugin-eval.md.
Layer 2 — Human prompt suite (required for material edits)
- Write 3–5 representative user prompts (should trigger + should not trigger).
- Run agent without skill → record failures.
- Run with skill → record improvements and new failures.
- Mirror prompts in
evals/cases/*.yaml(CI rules) andreferences/evals.md(behavior log).
Split ~60% train (edit against) / 40% held-out (gate acceptance)—mirrors SkillOpt selection gate.
Layer 3 — SkillOpt-inspired improve loop (research)
SkillOpt treats SKILL.md as trainable text with a frozen agent:
Rollout (tasks + current skill) → Reflect (failures vs successes)
→ Bounded edit (add/delete/replace under budget) → Held-out gate (keep only if better)Skill Steward manual adaptation (no GPU cluster required):
| Step | Action |
|---|---|
| 1 | Baseline: held-out pass rate without skill |
| 2 | With skill: same tasks, log pass rate |
| 3 | Reflect: list 1–3 concrete failure modes |
| 4 | Bounded edit: ≤10% line churn or one new section; no wholesale rewrite |
| 5 | Re-run held-out only; keep edit only if improved |
| 6 | Record outcome in references/evals.md + sources.md |
Paper: https://arxiv.org/abs/2605.23904 · Site: https://microsoft.github.io/SkillOpt/
Related: SkillLens (model-generated skills study).
Show full SKILL.md (336 more words)Show less
Layer 4 — Ecosystem benchmarks (2026+)
| Resource | Use |
|---|---|
| SkillsBench | Inspiration for paired vanilla vs skill-augmented tasks |
| skillgrade | Regression testing skill quality (mgechev) |
| Claude authoring best practices | Eval-before-write workflow |
Layer 5 — Runtime Dogfood Benchmarks
At 10,000x scale, NLP prompt evaluation fails because LLMs suffer Cognitive Overload navigating massive toolsets. Runtime dogfood should objectively assert their logical trajectory using deterministic traces.
- Capture compact traces: Store action IDs, tool counts, artifact digests, and redacted excerpts, not raw product traces.
- Define assertions: Expected action trajectory, declared surfaces used first, maximum tool calls, maximum repair/setup attempts, maximum unrelated tool calls, required
return_to_goal_step, required artifacts, and negative checks for unrelated actions. - Run dogfood benchmarks: Use
steward benchmark --scenario <id> --jsonfor runtime dogfood scenarios. Do not put product runtime scenarios under T1 behavior-critical skill evals.
The current steward eval --name registered-eval path is legacy/experimental. Skill quality remains pnpm run eval; runtime dogfood belongs to steward benchmark, where durability_blocked is valid blocked evidence when contract inputs are modified or untracked, not proof of runtime behavior.
Improve workflow (checklist)
- [ ] sources.md cites plugin-eval + SkillOpt if used
- [ ] pnpm run validate
- [ ] T1 behavior-critical: `pnpm run eval` + cases updated
- [ ] plugin-eval analyze (optional)
- [ ] 3+ prompt evals documented in references/evals.md
- [ ] Bounded edit applied; held-out improved
- [ ] skill-authoring-lifecycle checklist
- [ ] PR mentions eval deltaWhat to fix first (typical order)
name/description(routing)—must include what + when- Broken links / missing
references/sources.md - Delete or replace duplicated rules before adding a new section or eval case
- Move bulk to
references/(SKILL.md < 500 lines) - Add error-handling / validation steps agents skip
- Token cost (description length, always-loaded content)
Anti-patterns
- Rewriting entire SKILL.md from one failure (destroy working rules)
- Self-editing without held-out prompts (overfit)
- Adding skill rules, evals, or tools from one observed run when a smaller FAQ, error message, native command, observed-effect check, or deletion would solve the problem
- Adding a new eval for duplicated guidance before trying to compress, delete, or replace the overlapping rule
- Claims without
references/sources.mdrows - Evaluating only with static analyze—never running real prompts
- LLM judge in CI (flake, cost) — offline only per ADR 0011
- Passing
pnpm run evaland claiming agent behavior is proven
Related skills
| Skill | Role |
|---|---|
skill-source-citations | Save research links |
skill-authoring-lifecycle | Scaffold |
skill-authoring-lifecycle | Pre-merge audit |
Sources
Install
npx skills add arenukvern/skill_steward --skill skill-eval-improve© Arenukvern, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 11 other files (references) in .agents/skills/skill-eval-improve of Arenukvern/mcp_flutter.
- SKILL.md
- evals/cases/duplicated-guidance-compression-trigger.yaml
- evals/cases/improve-routing-trigger.yaml
- evals/cases/tool-loop-drift-trigger.yaml
- evals/cases/validate-only-dormant.yaml
- references/chrome-eval-design.md
- references/eval-case-schema.md
- references/evals-template.md
- references/evals.md
- references/plugin-eval.md
- references/skillopt-and-research.md
- references/sources.md
Open the folder on GitHubat commit 62f3ee1
Compare with similar skills
Skill Eval Improve next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Skill Eval Improve this skillArenukvern/mcp_flutter | 386 | — | ~2.4k | Automated safety check: Pass | MIT | |
| Aeon Skill EvalsBankrBot/skills | 1.2k | — | ~660 | Automated safety check: Pass | None | |
| Waza Skill Evaluatormicrosoft/waza | 1.4k | — | ~2k | Automated safety check: Pass | MIT | |
| Waza Interactivemicrosoft/waza | 1.4k | — | ~1.3k | Automated safety check: Pass | MIT | |
| Continuous Agent Loopaffaan-m/ECC | 276k | 5 repos | ~298 | Automated safety check: Pass | MIT | |
| Skill Release Gaterohitg00/ai-engineering-from-scratch | 66k | — | ~1k | Automated safety check: Pass | MIT |
Similar skills
- Official
Waza Skill Evaluator
microsoft/waza
Evaluates agent skills with a Go CLI that runs YAML-defined benchmarks, compares runs and scores the quality of SKILL.md frontmatter.
- Official
Waza Interactive
microsoft/waza
Walks you through creating, running and reading waza evals for an agent skill, then proposes concrete fixes when tasks fail or the score is low.
Continuous Agent Loop
affaan-m/ECC
Patterns for continuous autonomous agent loops with quality gates, evals, and recovery controls.
Skill Release Gate
rohitg00/ai-engineering-from-scratch
Evaluates an Agent Skill bundle before release for structure, trigger quality, artifact improvement, script correctness, safety, installed-tree integrity and host portability.
- Official
Operational Value Designer
github/gh-aw
Designs and verifies a deterministic grader that measures whether a GitHub Agentic Workflow run reached its real-world or repository outcome.
More from Arenukvern/mcp_flutter
All 23 skills in this repoHarness Engineering Lifecycle
Arenukvern/mcp_flutter
Design, implement, and integrate generalized validation harnesses across a producer-consumer boundary after a local harness contract exists.
Mixture Of Experts
Arenukvern/mcp_flutter
Run a Mixture of Experts (MoE) audit on any topic, plan, codebase, evidence archive, or process.
Multi Agent Handoff
Arenukvern/mcp_flutter
Plan and document handoffs, parent lane contracts, and parallel batch contracts between specialized AI agents (foreman, workers, reviewers).
Plugin Marketplace Setup
Arenukvern/mcp_flutter
Designs public or private Agent Skill and plugin marketplaces for Cursor, Claude Code, Codex, Zed, Open Plugin, and npx skills—manifest layout, install matrix, and Skill Steward vs product boundaries.
Release Changelog Harness
Arenukvern/mcp_flutter
Chooses ecosystem-native release and changelog tooling (Changesets, Melos, release-plz) plus binary distribution (GitHub Release tarballs, install.sh) when the product is an executable.
Repository Governance Lifecycle
Arenukvern/mcp_flutter
Master orchestration for repository governance, North Star impact, sub-Star boundaries, and repair-first or evidence-first drift checks.
Related
Works with
Categories
Questions about Skill Eval Improve
What does Skill Eval Improve do?
Improves Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with held-out gates. Skill Eval Improve is an agent skill from Arenukvern/mcp_flutter. Improves Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with held-out gates.
When should I use Skill Eval Improve?
Skill Eval Improve fits situations like: tuning skill quality; adopting Chrome/Microsoft T-named quality gates—not for bulk validate-only; skillOpt automation.
How do I install Skill Eval Improve in Claude Code?
Run `npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a claude-code`. Or copy the skill folder (.agents/skills/skill-eval-improve in Arenukvern/mcp_flutter) into .claude/skills/skill-eval-improve in your project. Claude Code loads it when a task matches its description.
How do I install Skill Eval Improve in Codex?
Run `npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a codex`. Or copy the skill folder (.agents/skills/skill-eval-improve in Arenukvern/mcp_flutter) into .agents/skills/skill-eval-improve in your project. Codex loads it when a task matches its description.
Can I use Skill Eval Improve in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Arenukvern/mcp_flutter --skill skill-eval-improve -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/skill-eval-improve, .gemini/skills/skill-eval-improve, .github/skills/skill-eval-improve and .opencode/skills/skill-eval-improve in your project.
What does Skill Eval Improve need to run?
Going by SKILL.md and its folder, Skill Eval Improve needs the command-line tools its instructions call (pnpm and npx). Our summary lists: Node.js.
Does Skill Eval Improve access the network?
SKILL.md names 3 domains. As links in the text: microsoft.github.io, arxiv.org and github.com. This is read from the text; nothing was executed.
Is Skill Eval Improve safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Skill Eval Improve use?
Skill Eval Improve is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Skill Eval Improve use?
About 2.4k tokens (SKILL.md is roughly 9.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 3k tokens, read only when the agent opens those files.
What are the alternatives to Skill Eval Improve?
Skills that share tags, products or a category with Skill Eval Improve: Aeon Skill Evals (BankrBot/skills, 1.2k stars), Waza Skill Evaluator (microsoft/waza, 1.4k stars), Waza Interactive (microsoft/waza, 1.4k stars) and Continuous Agent Loop (affaan-m/ECC, 276k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Skill Eval Improve?
Arenukvern (a GitHub user) maintains it in Arenukvern/mcp_flutter, which has 386 GitHub stars. The repository holds 23 skills in this directory. The repository was last updated on October 3, 2026.
Source: Arenukvern/mcp_flutter on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.