Is the Eval Consistency skill free?

Yes. The skill itself is free from YIKUAIBANZI/forge-skill. SkillProof publishes the install command and an independent test verdict at no cost.

What is the Eval Consistency SkillProof Score?

8.0/10 — installs cleanly 4/5, triggers reliably 5/5, output vs. baseline 7/10, docs & honesty 4/5.

How do I install Eval Consistency?

Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.

Can I use Eval Consistency with Cursor, Copilot, Gemini CLI, Codex or other AI tools?

The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.

Eval Consistency

Scores a persona's roleplay replies on 5 dimensions and reports a 0-100 number

by YIKUAIBANZI · YIKUAIBANZI/forge-skill

Works with setup

Install — copy the command Download source ↗

Test report

Verdict: Works with setup
Score: 8.0/10
Tested: Aug 5, 2026
Environment: Claude Code 2.x (agent harness)

⚠ This skill is no longer available upstream. Our re-check on Aug 10, 2026 couldn't find it any more (repo unreachable/deleted). The test below is what we measured on Aug 5, 2026 and we're leaving it up as a record — but there is nothing left to install, so we've removed the command.

Built a schema-valid persona.json for the profile the bundled test cases target and confirmed it with the repo's own persona_validator.py, then ran three scenarios twice, with and without the skill. The no-skill baseline produced a prose verdict ('c02 feels a little flat'); the skill produced 25 scored judgments, a comparable 96.7/100 aggregate, and pinpointed the actual cause, that the scenario's natural phrasing lives in L0 and L4 evidence but was never registered in L2.signature_phrases. That pointer at a specific JSON field is the real gain over the baseline. The weakness is that the same model both writes and grades the replies, so a 70-point pass bar barely discriminates. It also cannot run out of the box: personas/ is gitignored, so you must first generate a profile with /forge-persona, and the Step 0 paths only resolve from inside a forge-skill checkout.

Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works

Installs cleanly 4/5
Triggers reliably 5/5
Output vs. baseline 7/10
Docs & honesty 4/5

What Eval Consistency does

Runs a repeatable roleplay-consistency evaluation over a forge-persona profile: it replies to a fixed set of chat scenarios in character, then grades each reply on message length, signature phrases, punctuation, interaction pattern and hard boundaries. Output is a per-scenario breakdown, per-dimension averages and one comparable 0-100 score you can re-run after editing a persona. Triggers on /eval-consistency or on a request to test persona or roleplay consistency.

How to install Eval Consistency

Nothing to install: the source repository no longer has this skill. If the author brings it back, our daily re-check will pick it up and the command will reappear here.

Commands — how to trigger Eval Consistency

/eval-consistency Scores a persona's roleplay replies on 5 dimensions and reports a 0-100 number

It also activates on plain-language prompts like these:

Run /eval-consistency to score my persona's roleplay replies
Test whether my forge persona stays in character across scenarios
跑一下我这个角色扮演人设在这几个对话测试场景下的一致性评测，看看最终能打多少分数

Frequently asked questions

Is the Eval Consistency skill free?: Yes. The skill itself is free from YIKUAIBANZI/forge-skill. SkillProof publishes the install command and an independent test verdict at no cost.
Does Eval Consistency work with Claude Code?: We tested it with Claude Code 2.x (agent harness) on Aug 5, 2026. Verdict: Works with setup. Built a schema-valid persona.json for the profile the bundled test cases target and confirmed it with the repo's own persona_validator.py, then ran three scenarios twice, with and without the skill. The no-skill baseline produced a prose verdict ('c02 feels a little flat'); the skill produced 25 scored judgments, a comparable 96.7/100 aggregate, and pinpointed the actual cause, that the scenario's natural phrasing lives in L0 and L4 evidence but was never registered in L2.signature_phrases. That pointer at a specific JSON field is the real gain over the baseline. The weakness is that the same model both writes and grades the replies, so a 70-point pass bar barely discriminates. It also cannot run out of the box: personas/ is gitignored, so you must first generate a profile with /forge-persona, and the Step 0 paths only resolve from inside a forge-skill checkout.
What is the Eval Consistency SkillProof Score?: 8.0/10 — installs cleanly 4/5, triggers reliably 5/5, output vs. baseline 7/10, docs & honesty 4/5.
How do I install Eval Consistency?: Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
Can I use Eval Consistency with Cursor, Copilot, Gemini CLI, Codex or other AI tools?: The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.