Ops — does it run?
Brief 23

Release Pipeline Audit

Map merged-PR to production; every manual step is a finding. Includes the rollback-readiness verdict.

RELEASE.md4 phases7 lenses~2.5k chars

Free & open · no signup · read-only — it ends by asking · nothing leaves your machine

What it does

You are working inside this repo. Mission: map the full path from merged PR to production, treat every manual step as a finding, and deliver a verdict on rollback readiness.

Read-only pass: inspect CI configs, deploy scripts, docs. Your only write is the report file.

What it looks for

7 lenses it looks through

Phase 2 sweeps the codebase through every one of these, citing file and line for each finding.

1

CI health

duration, flaky steps, and the checks NOT run (types, lint, security scans, size budgets)

2

Manual steps

every human action between merge and production: each is latency, error surface, and a bus-factor risk

3

Rollback path

does one exist; has it ever been exercised; are data migrations reversible or one-way doors?

4

Environment parity

differences between local/staging/prod that produce works-on-my-machine failures

5

Migration safety

how schema changes deploy relative to code; the window where they can disagree

6

Flag hygiene

feature flags: kill switches for risky changes, and stale flags rotting in the code

7

Release visibility

can anyone tell what shipped, when, and what changed? Changelog, tags, deploy log

The deliverable

What lands in your repo

One structured report at the repo root — or in reports/, if you keep one — the same shape every time, ready for a teammate — or the optional Studio — to act on.

RELEASE.md
1
Pipeline map
stage · time · automated? · risk noted
2
Rollback verdict
ready / partial / one-way, with evidence
3
Findings
each: issue · risk · fix · effort
4
The one automation
the single step to automate first and its payoff
5
Deploy checklist
interim, until automation lands
One file. Evidence-backed. It ends by asking before touching anything.
How it works

What it makes the agent do

Every Goal Prompt follows the same four steps, so results are consistent and repeatable — no matter which one you run.

1
Phase 1

Map the pipeline

Trace the actual path.

2
Phase 2

Audit through 7 lenses

CI health — duration, flaky steps, and the checks NOT run (types, lint, security scans, size budgets).

3
Phase 3

Curate

Rank by deploy-confidence gained per unit effort.

4
Phase 4

Report

Create RELEASE.md at repo root.

Get started

Three ways to run this Goal Prompt

01 · COPY

Paste it in

Copy it and paste it into your agent inside the repo you want checked.

02 · INSTALL

As a slash command

Install the goal plugin once — two commands — then just type /goal:release-pipeline-audit.

1/plugin marketplace add GhostlyGawd/goal-prompts
2/plugin install goal@goal-prompts

Or install only this Goal Prompt as /goal-release-pipeline-audit:

curl -fsSL https://goal-prompts.vercel.app/install | BRIEF=23 sh
03 · AGENT

From an agent (MCP)

Let an agent fetch it mid-conversation, or pull the raw Goal Prompt by URL.

https://goal-prompts.vercel.app/raw/23.md
Transparency

The exact prompt

Nothing hidden — this is the whole Goal Prompt, verbatim. Read it in a minute, edit it, or copy it as-is.

Read the full brief (2,516 characters)
# Goal: Release Pipeline Audit

You are working inside this repo. Mission: map the full path from merged PR to production, treat every manual step as a finding, and deliver a verdict on rollback readiness.

Read-only pass: inspect CI configs, deploy scripts, docs. Your only write is the report file.

## Phase 1 — Map the pipeline
- Trace the actual path: merge → build → test → deploy → verify. From configs and scripts, not from memory or docs.
- Time each stage where measurable; note which steps are automated vs human-driven.
- Who can deploy, from where, and what does a deploy require?

## Phase 2 — Audit through 7 lenses
1. **CI health** — duration, flaky steps, and the checks NOT run (types, lint, security scans, size budgets)
2. **Manual steps** — every human action between merge and production: each is latency, error surface, and a bus-factor risk
3. **Rollback path** — does one exist; has it ever been exercised; are data migrations reversible or one-way doors?
4. **Environment parity** — differences between local/staging/prod that produce works-on-my-machine failures
5. **Migration safety** — how schema changes deploy relative to code; the window where they can disagree
6. **Flag hygiene** — feature flags: kill switches for risky changes, and stale flags rotting in the code
7. **Release visibility** — can anyone tell what shipped, when, and what changed? Changelog, tags, deploy log

## Phase 3 — Curate
- Rank by deploy-confidence gained per unit effort
- Every finding cites the config, script, or absence thereof

## Phase 4 — Report
Create `RELEASE.md` at repo root:
1. **Pipeline map** — stage · time · automated? · risk noted
2. **Rollback verdict** — ready / partial / one-way, with evidence
3. **Findings** — each: issue · risk · fix · effort
4. **The one automation** — the single step to automate first and its payoff
5. **Deploy checklist** — interim, until automation lands

Start the report with today's date. If `RELEASE.md` already exists from a previous run, read it first and lead with what changed since.

## Rules
- Judge the pipeline that exists, not the one described in docs
- Boring, frequent, reversible deploys are the goal
- No release pipeline in this repo? Say so in a one-paragraph null report and stop — a null result is a valid finding.
- If a `reports/` directory exists at the repo root, write the report there instead of the root.
- Before asking, present the top findings as a ranked list in plain words
- Report only — end by asking which fixes to make

House rules for this Goal Prompt

  • Judge the pipeline that exists, not the one described in docs
  • Boring, frequent, reversible deploys are the goal
  • No release pipeline in this repo? Say so in a one-paragraph null report and stop — a null result is a valid finding.
  • If a reports/ directory exists at the repo root, write the report there instead of the root.
  • Before asking, present the top findings as a ranked list in plain words
  • Report only — end by asking which fixes to make
Keep exploring

More Ops briefs

Runs inside these playbooks — curated sequences you can launch with one paste: