ai-development-guide

A guide for investigating bugs and implementation problems by tracing them back to their earliest incorrect assumption or contract.

In plain words
What is it for?
It supports bug fixes, refactoring, code reviews, and quality checks by comparing expected and actual behavior, inspecting affected code, and adding focused regression checks.
Why use it?
It helps avoid piling workarounds onto the wrong part of a system and keeps fixes within the approved scope.

Skill for Claude CodeCodex

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/shinpr/agentic-code/ai-development-guide
Any agent
npx skills add shinpr/agentic-code --skill ai-development-guide
Clone the repo
git clone --depth 1 https://github.com/shinpr/agentic-code

Made for: Claude Code, Codex.

Per session 39 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 595 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 $0.00039 $0.00595
Opus 5 $0.00019 $0.00298
Sonnet 5 $0.00008 $0.00119
Haiku 4.5 $0.00004 $0.00060

Measured 2d ago against content hash 70a9832b8d52, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ai-development-guide 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 2d 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.

.agents/skills/ai-development-guide/SKILL.md · 74 lines

How it starts

The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AI Development Guide

Core Rule

Correct the invalid source before adding compensating layers. A workaround, fallback, abstraction, or new artifact is justified only when current evidence shows that removing or fixing the cause cannot satisfy the approved outcome.

Investigation

Inspect the smallest relevant path that can establish:

  • expected versus actual behavior;
  • the first failing boundary;
  • callers, contracts, and data or control flow affected;
  • existing equivalent behavior and representative repository patterns;
  • tests and checks that can reproduce or observe the issue;
  • unknowns that could change the fix.

Mark findings as observed, inferred, or unknown. Expand inspection when new evidence points to another affected boundary, not to satisfy a fixed coverage checklist.

Root-Cause Resolution

  1. Reproduce or otherwise observe the failure.
  2. Identify the earliest invalid assumption, state, or contract.
  3. Determine whether existing code, configuration, or documentation already provides the correct mechanism.
  4. Fix or remove the invalid source.
  5. Add a focused regression proof when it protects the demonstrated failure from recurrence.

Use a minimal reproduction or iterative why-analysis when it changes the cause decision. Do not create either as a ceremonial artifact.

Existing Mechanism Decision

When similar functionality exists:

  • reuse it when its responsibility, lifecycle, and contract fit the current outcome;
  • extend it when the extension preserves its responsibility and remains cheaper than a parallel mechanism;
  • replace or remove it when evidence shows it is broken or unused;
  • create a new mechanism only when the current ones cannot satisfy the approved requirement.

Technical debt discovered outside the current outcome remains a finding, not an automatic ADR or implementation task.

Fallbacks

Prefer explicit failure with preserved error context. Add fallback behavior only when a current requirement defines the degraded outcome or evidence shows continuity is necessary. Verify that the fallback is observable and does not hide the original failure.

Read the full file on GitHub · 74 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. 2d ago First seen · 74 lines · 39 tokens per session scan A 70a9832b8d52

Subscribe to this mod's changes

ai-development-guide is a skill published in the GitHub repository shinpr/agentic-code (49 stars, last pushed 3d ago), licensed MIT. It adds 39 tokens to every session and 595 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 2026-08-30.