ai-coding-workflow

A set of workflow rules for scoped software development, independent quality checks, and human approval steps.

In plain words
What is it for?
Use it to manage development slices, maintain workflow documents, require independent QA, and prevent implementation or deployment before the specified approvals.
Why use it?
It keeps planning, implementation, testing, and release decisions in defined phases with required approval gates.

Cursor rule for Cursor

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 rules/espenandreass1/agentslice/ai-coding-workflow
Clone the repo
git clone --depth 1 https://github.com/Espenandreass1/agentslice

Made for: Cursor.

Per session 411 This file is loaded in full into every session.
When invoked 411 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 $0.00411 $0.00411
Opus 5 $0.00205 $0.00205
Sonnet 5 $0.00082 $0.00082
Haiku 4.5 $0.00041 $0.00041

Measured 2d ago against content hash 91e52994cbf5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ai-coding-workflow 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.

kit/.cursor/rules/ai-coding-workflow.mdc · 18 lines

What it actually says

AgentSlice Rules

Follow AGENT_RULES.md; docs/planning/workflow-state.md alone controls active phase and approval fields.

At every start, resume, or phase change, read AGENT_RULES.md, docs/planning/active-context.md, docs/planning/workflow-state.md, docs/planning/current-slice.md, and docs/planning/checkpoint.md. The checkpoint is a handoff summary, never state truth. Then add only role context: planning reads vision, tech stack, live next slices; build reads approved spec and coding rules; QA reads approved spec, changed files, QA plan, plus one relevant prior QA report only when needed; release reads current spec, QA report, and changelog.

Do not bulk-read historical folders. Use docs/archive/README.md plus targeted search only for a concrete historical need.

Keep the four gates unchanged: human slice approval, human spec approval, independent QA PASS/PASS WITH NOTES, then human release approval. Never implement without the first two, recommend release after FAIL, or deploy without the last. Quick fixes remain fully gated and never reduce auth, ownership, data, privacy, commerce, shared-contract, database/RLS, migration, or production safeguards.

Keep active documents compact: active-context.md ≤180 lines; checkpoint.md ≤80 lines; next-slices.md has one to three living candidates; decisions.md has active constraints only; changelog entries are short and user-facing. At handoff refresh the checkpoint; after a human-approved gate archive only superseded history indexed by docs/archive/README.md. QA defaults to focused acceptance/domain regression and must state any full-suite trigger. Do not parallelize by default; independent QA/review is the normal exception. Keep output compact and link to earlier evidence instead of repeating it.

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 · 18 lines · 411 tokens per session scan A 91e52994cbf5

Subscribe to this mod's changes

ai-coding-workflow is a cursor rule published in the GitHub repository Espenandreass1/agentslice (4 stars, last pushed 7d ago), licensed MIT. It adds 411 tokens to every session, about $0.0021 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.