skill-forge

A quality-checking process for agent instruction files such as SKILL.md, CLAUDE.md, AGENTS.md, and Cursor rules. It tests a draft with adversarial prompts and uses judges to score and refine it before promotion.

In plain words
What is it for?
It hardens skills and other agent instruction files through repeated review rounds. It can adapt its tests to the type of document being checked and stop at a budget limit if the draft does not pass.
Why use it?
It helps expose instructions that can be misled, misunderstood, or bypassed by prompt injection. The process provides a quality gate after an instruction file has been written.

Skill for Claude CodeCodex

Part of the ai-native-toolkit plugin — 15 skills, 7 commands, 8 agents shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/bjcoombs/ai-native-toolkit/skill-forge
Any agent
npx skills add bjcoombs/ai-native-toolkit --skill skill-forge
Clone the repo
git clone --depth 1 https://github.com/bjcoombs/ai-native-toolkit

Made for: Claude Code, Codex.

Or install ai-native-toolkit, the plugin that ships this one along with the rest of its 15 skills, 7 commands, 8 agents.

Per session 156 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 6,140 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 $0.00156 $0.06140
Opus 5 $0.00078 $0.03070
Sonnet 5 $0.00031 $0.01228
Haiku 4.5 $0.00016 $0.00614

Measured 3d ago against content hash 4f61fad9dd48, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skill-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 3d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (tests/fixtures/flawed-instruction-file/check_counts.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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/skill-forge/SKILL.md · 176 lines

How it starts

The opening of the file, as written. The whole thing — 176 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Skill-forge - judge-panel skill-hardening harness

A lead (the Forge Master) chairs ephemeral runner teammates that exercise a draft document via prompt-injection, and a persistent judge panel that scores the transcripts. The lead amends one thing per round and applies a strict 3-tier promotion gate, until the document promotes or hits a budget ceiling. It is proven by forging itself.

The target is a SKILL.md by default, but the same loop hardens any agent instruction file - CLAUDE.md, AGENTS.md, GEMINI.md, .cursor/rules/*, .github/copilot-instructions.md - because those files have the same failure surface as a skill: they steer an agent and can mislead it. The artifact type is detected deterministically and drives three adaptations (the lens set, the runner variant, and the canonical intent); see Artifact type and the lens-selector below.

Boundary

Skill-forge is a prove-and-promote quality gate that runs after authoring, not an authoring tool: it pairs with authoring skills (brainstorming, superpowers writing-skills) that produce a draft, then drives that draft through adversarial rounds before it ships. For an instruction file the same boundary holds - it gates a CLAUDE.md/AGENTS.md that already exists or has just been drafted; it does not write one from scratch.

Input contract

Input Required Notes
Target yes Path to a SKILL.md (default), or to any agent instruction file (CLAUDE.md, AGENTS.md, GEMINI.md, .cursor/rules/*, .github/copilot-instructions.md), or inline draft text. The artifact type is detected from the path/filename - see Artifact type and the lens-selector.
Intent / spec yes What it should do and who/what should trigger it. The ground truth the Fidelity lens judges against. For an instruction file the canonical intent is derivable - see the derivation guard below.
Known failure modes / existing tests no Seeds the test corpus.

Read the full file on GitHub · 176 lines

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. 3d ago First seen · 176 lines · 156 tokens per session scan A 4f61fad9dd48

Subscribe to this mod's changes

skill-forge is a skill published in the GitHub repository bjcoombs/ai-native-toolkit (30 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 156 tokens to every session and 6,140 once invoked, about $0.0008 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 2026-08-30.