forge

forge is a skill for Claude Code from quantsquirrel/claude-forge-smith. It costs 37 tokens per session (3,425 once invoked), scanned A, original, MIT.

A system for improving a skill through test-driven evaluation, where TDD means testing behaviour before and after making a change. It uses repeated evaluations and statistical checks to assess whether the skill improved.

In plain words
What is it for?
It is for testing a skill against challenging scenarios, finding weaknesses, applying upgrades, and checking the results before keeping them.
Why use it?
It helps distinguish a real improvement from a change that only appears better in one trial.

Skill for Claude Code

Written for Claude Code: allowed-tools in frontmatter. Also seen: reads .claude/ paths; mentions subagents.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the forge plugin — 3 skills, 2 hooks shipped together

Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add quantsquirrel/claude-forge-smith
Claude Code
/plugin install forge

Made for: Claude Code.

Or install forge, the plugin that ships this one along with the rest of its 3 skills, 2 hooks.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for forge

README.md
[![agentmods](https://agentmods.dev/badge/skills/quantsquirrel/claude-forge-smith/forge.svg)](https://agentmods.dev/skills/quantsquirrel/claude-forge-smith/forge)
Your own site
<a href="https://agentmods.dev/skills/quantsquirrel/claude-forge-smith/forge"><img src="https://agentmods.dev/badge/skills/quantsquirrel/claude-forge-smith/forge.svg" alt="Measured on agentmods" height="20"></a>
Per session 37 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,425 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00037 $0.03425
Opus 5 $0.00018 $0.01713
Sonnet 5 $0.00007 $0.00685
Haiku 4.5 $0.00004 $0.00343

Measured 5d ago against content hash 79c7e225ae35, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

forge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/forge/SKILL.md · 429 lines

How it starts

The opening of the file, as written. The whole thing — 429 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Forge Skill

Systematically improve a skill through test-driven evaluation and statistical validation.

REQUIRED BACKGROUND: First invoke forge:smelt skill for TDD methodology applied to skills. See testing-skills-with-subagents.md in that skill's directory for pressure scenario templates.

Quick Reference

Step Action
1 Trial Branch 생성
2 스킬 찾기 및 읽기
3 Pressure Scenario 생성
4 기준선 평가 (3회) - evaluator 사용
5 Discoverability 평가 (CSO 체크)
6 개선 사항 식별
7 개선 적용 (GREEN Phase)
8 개선 후 평가 (3회)
9 신뢰구간 분리 확인
10 Trial Branch 결과 처리 (병합/폐기)
11 Stats 업데이트 (upgraded: true)

Upgrade Mode Selection

스킬 업그레이드 시작 전, 적합한 모드를 자동 선택합니다.

Mode Decision Flow

스킬 분석 (get_upgrade_mode 호출)
    │
    ├─ "TDD_FIT" ──→ TDD Mode (기존 워크플로우)
    │                 - Trial Branch
    │                 - 3x 평가 + 95% CI (또는 n=5 고정밀)
    │                 - 통계적 검증
    │
    └─ "HEURISTIC" ──→ Heuristic Mode (신규)
                       - Usage 데이터 분석
                       - 구조 품질 평가
                       - 자동 개선 제안

Mode Detection

업그레이드 시작 시 다음 bash 함수를 호출하여 모드 결정:

source "${CLAUDE_PLUGIN_ROOT}/hooks/lib/storage-local.sh"
MODE=$(get_upgrade_mode "$skill_name")

TDD Mode Options

Option Sample Size CI Width When to Use
Standard n=3 Wider 빠른 피드백, 대부분의 경우
High Precision n=5 Narrower 미묘한 개선 검증, 중요한 스킬

사용자가 /forge --precision=high 또는 /forge -n5로 n=5 모드 선택 가능

TDD Mode (기존)

  • 조건: 테스트 파일 또는 pressure-scenarios.md 존재
  • 검증: check_skill_has_test() → true
  • 워크플로우: Step 1-11 (기존 그대로)

Heuristic Mode (신규)

  • 조건: 테스트 파일 없음
  • 검증: get_upgrade_mode() → "HEURISTIC"
  • 워크플로우:
    1. Usage 데이터 로드 (get_all_skills_summary())
    2. 서브에이전트 호출: Task(subagent_type="forge:heuristic-evaluator", prompt="Evaluate skill: <skill-name>")
    3. 점수 60 미만 → 자동 개선 제안 적용
    4. Trial Branch에서 개선 적용
    5. 1주일 후 사용량 변화로 검증 (get_usage_trend())

Read the full file on GitHub · 429 lines

Files

What ships with it

4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 429 lines · 0 tokens per session scan A 79c7e225ae35

Subscribe to this mod's changes

forge is a skill published in the GitHub repository quantsquirrel/claude-forge-smith (2 stars, last pushed 6mo ago), licensed MIT. It adds 37 tokens to every session and 3,425 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.