delegate-and-verify

delegate-and-verify is a skill for Claude Code, Codex from benjaminard/fable-skills. It costs 70 tokens per session (637 once invoked), scanned A, original, MIT.

A working method for splitting independent coding tasks among helper agents and checking the results with a fresh review.

In plain words
What is it for?
Use it for multi-part changes, audits, migrations, reviews, and other work that benefits from parallel tasks and separate verification.
Why use it?
It reduces serial work and helps catch mistakes that are easy to miss when reviewing your own changes.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: mentions subagents.

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/benjaminard/fable-skills/delegate-and-verify
Any agent
npx skills add benjaminard/fable-skills --skill delegate-and-verify
Clone the repo
git clone --depth 1 https://github.com/benjaminard/fable-skills

Made for: Claude Code, Codex.

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 delegate-and-verify

README.md
[![agentmods](https://agentmods.dev/badge/skills/benjaminard/fable-skills/delegate-and-verify.svg)](https://agentmods.dev/skills/benjaminard/fable-skills/delegate-and-verify)
Your own site
<a href="https://agentmods.dev/skills/benjaminard/fable-skills/delegate-and-verify"><img src="https://agentmods.dev/badge/skills/benjaminard/fable-skills/delegate-and-verify.svg" alt="Measured on agentmods" height="20"></a>
Per session 70 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 637 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 $0.00070 $0.00637
Opus 5 $0.00035 $0.00318
Sonnet 5 $0.00014 $0.00127
Haiku 4.5 $0.00007 $0.00064

Measured 6d ago against content hash 9d70a751ec01, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

delegate-and-verify 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 6d 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/delegate-and-verify/SKILL.md · 29 lines

How it starts

The opening of the file, as written. The whole thing — 29 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Delegate and Verify

Two disciplines that compound: fan independent work out to subagents instead of doing everything serially, and verify finished work with a fresh context instead of your own re-reading.

Delegation

  1. Spot independence early. Before starting a multi-part task, ask: which parts share no state? Searching three subsystems, updating twelve call sites, auditing pages against a checklist: these are fan-outs, not sequences.
  2. Delegate and keep working. Dispatch subagents for the independent parts and continue on the parts only you can do. Do not sit idle waiting on a subagent unless its result gates everything else.
  3. Brief like a handoff. A subagent knows nothing you do not tell it. Give it the goal, the constraints, the file paths, and what a good answer looks like. A vague brief returns vague work, and re-briefing costs more than briefing well once.
  4. Intervene on drift. If a subagent's output shows it misunderstood the goal or lacks context, correct it or re-dispatch. Do not silently absorb wrong work into the result.

Verification

Fresh-context verifiers outperform self-critique. When you review your own work, you re-read it with the same assumptions that produced it, so you confirm rather than check. A verifier that never saw your reasoning has no such blind spot.

  1. Give the verifier the spec and the artifact, never your reasoning. It gets the original requirement and the finished work. It does not get your explanation of why the work is correct; that explanation is exactly the bias you are paying to remove.
  2. Ask it to break the work, not bless it. "Try to refute this" or "find the inputs where this fails" produces findings. "Does this look right?" produces agreement.
  3. On long autonomous runs, verify on an interval. Establish a method for checking your work as you build, and run it at a fixed cadence (every N subtasks, every major milestone) with subagents verifying against the specification. Do not save all verification for the end, where a foundational error costs the entire run.
  4. A clean review of risky work is itself a flag. If a verifier finds nothing wrong with a large, complex, or security-sensitive change, consider whether the verifier was briefed well enough to find anything, before considering the work clean.

Read the full file on GitHub · 29 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. 6d ago First seen · 29 lines · 70 tokens per session scan A 9d70a751ec01

Subscribe to this mod's changes

delegate-and-verify is a skill published in the GitHub repository benjaminard/fable-skills (30 stars, last pushed 2mo ago), licensed MIT. It adds 70 tokens to every session and 637 once invoked, about $0.0003 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.

Related

Other skills, from other repositories

guidance

Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidance - Microsoft Research's constrained generation framework.

davila7/claude-code-templates · 38 tokens

alterlab-paper-writer

Drafts and revises academic papers through a 12-agent pipeline with hardened LaTeX output (apa7 document class, justified text, table column-width formula, centered bilingual abstracts, standardized font stack, PDF compiled from LaTeX), supporting IMRaD, literature review, theoretical, case study, policy brief, and…

AlterLab-IEU/AlterLab-Academic-Skills · 276 tokens

alterlab-imaging-data-commons

Query and download public cancer imaging data from the NCI Imaging Data Commons (IDC) using the idc-index Python package, filtering by metadata, visualizing in-browser, and checking licenses, with no authentication required. Use when obtaining large-scale radiology (CT, MR, PET) or digital pathology DICOM datasets for…

AlterLab-IEU/AlterLab-Academic-Skills · 90 tokens

alterlab-molecular-dynamics

Runs and analyzes molecular dynamics simulations with OpenMM and MDAnalysis — setting up protein and small-molecule systems, assigning force fields, running energy minimization and production MD, and analyzing trajectories (RMSD, RMSF, contact maps, free energy surfaces). Use when simulating protein or ligand…

AlterLab-IEU/AlterLab-Academic-Skills · 98 tokens

alterlab-pyhealth

Develops, tests, and deploys clinical machine learning models with the PyHealth healthcare AI toolkit. Use when working with electronic health records (EHR), clinical prediction tasks (mortality, readmission, drug recommendation), medical coding systems (ICD, NDC, ATC), physiological signals (EEG, ECG), healthcare…

AlterLab-IEU/AlterLab-Academic-Skills · 117 tokens

alterlab-pufferlib

Scales reinforcement learning with PufferLib — high-throughput parallel training (PuffeRL), vectorized environments, and native multi-agent systems achieving 2-10x speedups over standard implementations. Use when scaling RL to millions of steps per second, running vectorized or multi-agent setups, building custom…

AlterLab-IEU/AlterLab-Academic-Skills · 127 tokens