Task: Remove gates for fully shipped features

SkillDev tools

Autonomous maintenance routine that finds gates for fully shipped

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Task: Remove gates for fully shipped features skill

What this skill tells your AI

The instructions your AI receives, as published by bex-co/beancount-io in .agents/skills/routine-shipped-feature-inliner/SKILL.md and read by ahel’s review.

Usage: /routine-shipped-feature-inliner [package-or-path]

Find a flag whose decision has already been made everywhere, prove it is forgotten rather than deliberate, inline the path it always takes, delete the other branch and the flag definition, and ship via /ship.

Contract

Read .agents/skills/routine-shared/contract.md first — preconditions, scope resolution, verify gates, ship protocol, budget, universal STOPs. Floor even without it: never ship red; one finding per ship; anything in a DO_NOT_DO.md is a hard STOP.

Workflow

1. Scope

Resolve scope per the contract.

2. Discover

Grep the package for gate shapes:

  • mobile/src/config.tsconfig.features.* compile-time constants
  • process.env.* used as a boolean switch; values seeded in .env.example and deploy configs (deploy/, root bex.yaml)
  • is*Enabled / enable* booleans, route guards, early returns on a config value

3. Prove it — deliberate gate vs forgotten flag (mandatory triage)

A flag is only a finding when all of these hold:

  1. It holds one value in every environment: grep every setter and reader, .env.example, deploy configs — no environment or code path flips it.
  2. Nothing marks it deliberate: check root .pm/DO_NOT_DO.md, the package's .pm/DO_NOT_DO.md, docs/adrs/, and code comments at the definition.
  3. git log on its introduction and last value change shows it has sat at its final value for a while, with no sign of a pending decision.

Any evidence of intent → not a finding; note it as skipped.

4. Fix

Inline the path the flag always takes. Delete: the now-dead branch, the flag definition, its plumbing (props/params that only carried it), and tests that only exercised the removed branch. Update tests that asserted the flag's value to assert the now-unconditional behavior. Never change the flag's value on the way out — inline only flags already at their final value.

5. Verify

Owning package's full gate per the contract. If the flag guarded UI, verify the surviving path still renders (mobile: light and dark per mobile/CLAUDE.md).

6. Ship

Compose /ship for this one flag. Loop within budget.

What NOT to do

  • NEVER inline, remove, or flip config.features.agentChat. The mobile agent-chat surface stays gated off by owner decision (mobile/.pm/DO_NOT_DO.md) — it is a deliberate gate, permanently out of scope.
  • The backend featureFlags GraphQL field is a cross-package contract (backend-cluster/backend-v2/src/features/ledger/api/resolvers/ledger-legacy-resolver.query.ts, consumed by the mobile app). Removing or reshaping it is a coordinated multi-package change — report it, don't fix it (universal STOP).
  • Never delete unflagged dead code — that is routine-dead-code-removal's job.
  • Never "finish shipping" a feature by turning its flag on. Flipping a value is a product decision, not maintenance.

Arguments

$ARGUMENTS

Signals

GitHub stars
267
Forks
30
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
routine-shipped-feature-inliner
Source
github.com/bex-co/beancount-io