meta-prompting

meta-prompting is a skill for Claude Code from jonathanung/finesse. It costs 15 tokens per session (4,468 once invoked), scanned A, original, MIT.

Guidance for writing prompts for a ralph loop, where an autonomous agent rereads the prompt from the beginning on every iteration. It defines required rules for measurable completion and recovery from failure.

In plain words
What is it for?
Creating Finesse prompts with command-checkable requirements, an explicit completion message, and instructions for handling repeated errors or blocked progress.
Why use it?
It reduces repeated unproductive attempts by making the target, stopping signal, checks, and fallback behavior explicit.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: mentions subagents; positional $N argument.

Part of the finesse plugin — 6 skills, 13 commands, 2 hooks shipped together

Good fit Creating Finesse prompts with command-checkable requirements, an explicit completion message, and instructions for handling repeated errors or blocked progress.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/jonathanung/finesse/meta-prompting
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.

Any agent
npx skills add jonathanung/finesse --skill meta-prompting
Clone the repo
git clone --depth 1 https://github.com/jonathanung/finesse

Made for: Claude Code.

Or install finesse, the plugin that ships this one along with the rest of its 6 skills, 13 commands, 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 meta-prompting

README.md
[![agentmods](https://agentmods.dev/badge/skills/jonathanung/finesse/meta-prompting.svg)](https://agentmods.dev/skills/jonathanung/finesse/meta-prompting)
Your own site
<a href="https://agentmods.dev/skills/jonathanung/finesse/meta-prompting"><img src="https://agentmods.dev/badge/skills/jonathanung/finesse/meta-prompting.svg" alt="Measured on agentmods" height="20"></a>
Per session 15 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,468 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 $0.00015 $0.04468
Opus 5 $0.00008 $0.02234
Sonnet 5 $0.00003 $0.00894
Haiku 4.5 $0.00002 $0.00447

Measured 8d ago against content hash 6c62f37ecdbf, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

meta-prompting 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 8d 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.

plugins/finesse/skills/meta-prompting/SKILL.md · 408 lines

How it starts

The opening of the file, as written. The whole thing — 408 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Ralph-Loop Prompt Construction

When drafting a ralph-loop prompt via Finesse, follow these rules. A ralph-loop prompt is NOT a conversation — it's a specification for an autonomous agent that re-reads it from scratch on every iteration with zero memory.

The 10 Mandatory Attributes

Every prompt you produce MUST have all 10. If any are missing, the validation agents will reject it.

1. Binary Completion Criteria

Every requirement must be checkable by running a command. If the agent can't programmatically verify success, it can't know when to stop.

  • YES: "All tests pass", "Coverage > 80%", "No linter errors", "API returns 200 for GET /users"
  • NO: "Code is clean", "Make it good", "Properly structured"

2. Explicit Completion Signal

The prompt must state exactly what to output and when: "Output <promise>COMPLETE</promise> when ALL of the following are true: [list]." Never leave this implicit.

3. Self-Diagnosing Failure Instructions

The agent WILL get stuck. The prompt must say what to do when stuck: "If stuck on the same error for 3+ attempts, step back, read the error trace fully, and try an alternative approach. If blocked for N+ iterations, document what's blocking progress and output <promise>BLOCKED</promise>."

4. Ordered Phases

"Build auth, then products, then cart" gives clear progression. A flat list of 20 unordered requirements is ambiguous about what to tackle next when the agent has no memory of what it was planning.

5. Verification Commands Baked In

Don't assume the agent will figure out how to check its work: "After each change, run npm test and npm run lint. If either fails, fix before moving on."

6. Guardrails Against Failure Modes

Explicit "Do NOT" rules based on known pitfalls:

  • "Do NOT rewrite files from scratch. Make targeted edits."
  • "Do NOT delete existing tests to make the suite pass."
  • "Read the actual error message before attempting a fix."
  • "Do NOT add unnecessary abstractions or extra files."

Note: Git commit and push rules are configured per-plan during Plan Construction — see Git Configuration Rules section below.

Read the full file on GitHub · 408 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. 8d ago First seen · 408 lines · 15 tokens per session scan A 6c62f37ecdbf

Subscribe to this mod's changes

meta-prompting is a skill published in the GitHub repository jonathanung/finesse (4 stars, last pushed 6mo ago), licensed MIT. It adds 15 tokens to every session and 4,468 once invoked, about $0.0001 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.

Related

Other skills, from other repositories

llm-app-patterns

Production-ready patterns for building LLM applications. Covers RAG pipelines, agent architectures, prompt IDEs, and LLMOps monitoring. Use when designing AI applications, implementing RAG, building agents, or setting up LLM observability.

davila7/claude-code-templates · 54 tokens

prompt-optimization

Improve a prompt on the evaluations workbench through a measured loop. Score the baseline first, then duplicate the target column, form a hypothesis from failing rows, edit the copy's prompt draft, run, compare pass rate and cost, and repeat until the numbers hold. Use when the user asks to optimize or improve a…

langwatch/langwatch · 105 tokens

enhance-prompt

Transforms vague UI ideas into polished, Stitch-optimized prompts. Enhances specificity, adds UI/UX keywords, injects design system context, and structures output for better generation results.

google-labs-code/stitch-skills · 41 tokens

prompt-engineer

Writes, refactors, and evaluates prompts for LLMs — generating optimized prompt templates, structured output schemas, evaluation rubrics, and test suites. Use when designing prompts for new LLM applications, refactoring existing prompts for better accuracy or token efficiency, implementing chain-of-thought or few-shot…

Jeffallan/claude-skills · 93 tokens

seedance-vocab-en

This skill should be used when an English Seedance 2.0 prompt needs clearer production wording, less generic prose, or precise vocabulary for camera, lighting, motion, VFX, audio, and constraints. Route blocked prompts through seedance-filter for context and boundary review.

Emily2040/seedance-2.0 · 61 tokens

ideogram4

Prompting patterns for Ideogram 4 text-to-image — best-in-class in-image text rendering and exact color/layout control via structured JSON captions. Use when generating images that need legible on-image text (title cards, thumbnails, logos, signage, CTAs), precise brand colors, or controlled spatial layout. Triggers…

digitalsamba/claude-code-video-toolkit · 99 tokens