agentic-sdlc-kit AGENTS.md

A set of instructions for building software with AI agents from an initial idea through review. It defines a lightweight engineering contract focused on evidence, clear decisions, small changes, and observable success.

In plain words
What is it for?
It helps inspect code, plan and implement changes, handle ambiguous requirements, run relevant checks, and report whether the requested result was actually achieved.
Why use it?
It reduces accidental scope expansion, hidden assumptions, unrelated refactoring, and claims that are not backed by verification. It also gives a clear process for asking the human to resolve important choices.

Instructions file for CodexOpenCode

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 instructions/rmjosea/agentic-sdlc-kit/agents-md
Clone the repo
git clone --depth 1 https://github.com/rmjosea/agentic-sdlc-kit

Made for: Codex, OpenCode.

Per session 1,539 This file is loaded in full into every session.
When invoked 1,539 The same file — it is already loaded in full.
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 $0.01539 $0.01539
Opus 5 $0.00770 $0.00770
Sonnet 5 $0.00308 $0.00308
Haiku 4.5 $0.00154 $0.00154

Measured 2d ago against content hash b4d5f7d81c26, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agentic-sdlc-kit AGENTS.md 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.md · 163 lines

How it starts

The opening of the file, as written. The whole thing — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent Engineering Contract

Purpose

Use this repository's skills to move from an idea to reviewed software with the least process that still produces reliable results.

Core behavior

  • Think before editing. Inspect relevant code and evidence first.
  • Surface material assumptions; never silently decide an ambiguous requirement.
  • Prefer the simplest solution that fully meets the accepted need.
  • Do not add unrequested features or speculative extensibility.
  • Make surgical changes. Do not refactor unrelated code.
  • Find root causes. Do not hide failures with workarounds.
  • Define observable success before implementation.
  • Run the relevant verification and report actual results.
  • Never claim completion when required checks are failing or were not run.
  • Preserve user changes and stop when scope or ownership is unclear.

Human decision protocol

Apply this protocol whenever a human decision is materially required.

  1. Ask exactly one question at a time.
  2. Explain the decision in plain, self-contained language.
  3. Do not combine independent decisions in one question.
  4. Offer two to four genuinely distinct options when useful.
  5. State the main consequence or trade-off of each option.
  6. Mark one option as Recommended and explain why it best fits the evidence.
  7. Allow a free-form answer when the listed options do not fit.
  8. After related decisions, summarize what is confirmed before continuing.

Ask when the answer could materially change scope, behavior, architecture, security, cost, user experience, reversibility, or the implementation plan. Do not interrupt for a trivial, reversible, repository-standard decision. Record such decisions with a concise rationale instead.

Workflow

The workflow is a set of composable gates, not a mandatory pipeline. Use the least durable process that safely preserves intent, evidence, and recoverability.

idea -> shape-idea -> write-spec -> plan -> decompose-tasks -> implement -> review
  • shape-idea: clarify conversationally; write no files.
  • write-spec: persist approved product behavior as one or more cohesive specs.
  • plan: resolve technical decisions and define the implementation design.
  • decompose-tasks: turn an approved plan into an executable dependency graph.
  • implement: complete one bounded direct, artifact-backed, or task-backed outcome using short feedback loops.
  • review: assess intent conformance and engineering quality independently.
  • create-skill: create or improve skills without bloating agent context.

Read the full file on GitHub · 163 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 · 163 lines · 1,539 tokens per session scan A b4d5f7d81c26

Subscribe to this mod's changes

agentic-sdlc-kit AGENTS.md is an instructions file published in the GitHub repository rmjosea/agentic-sdlc-kit (2 stars, last pushed 1mo ago), licensed MIT. It adds 1,539 tokens to every session, about $0.0077 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.