laicluse/agent-fieldkit - Page 2

install-hooks

Install git-discipline git hooks in the current repo so CLI/IDE commits get the same guardrails.

install-hooks

Install git-discipline git hooks in the current repo so CLI/IDE commits get the same guardrails.

how-plugins-work

Build, scaffold, or diagnose Claude/Codex plugins, skills, and marketplaces. Read this BEFORE creating a new plugin, marketplace, marketplace.json source, skill, or Codex adapter, and when generating adapters with Circus, not only when debugging plugin/skill…

test-before-push

Verify plugin candidates and integrate them without automatically mutating shared agent runtimes.

how-plugins-work

Build, scaffold, or diagnose Claude/Codex plugins, skills, and marketplaces. Read this BEFORE creating a new plugin, marketplace, marketplace.json source, skill, or Codex adapter, and when generating adapters with Circus, not only when debugging plugin/skill…

test-before-push

Verify plugin candidates and integrate them without automatically mutating shared agent runtimes.

testing-philosophy

Use when writing tests, debugging failures, reviewing test strategy, or judging suite health.

testing-philosophy

Use when writing tests, debugging failures, reviewing test strategy, or judging suite health.

circus

Sync multi-agent doctrine, AGENTS/CIRCUS files, and plugin metadata across Claude, Codex, and future agents.

circus

Sync multi-agent doctrine, AGENTS/CIRCUS files, and plugin metadata across Claude, Codex, and future agents.

rebase-latest-default

Rebase and verify the current worktree branch on the actual default-branch tip without using a default checkout.

rebase-latest-default

Rebase and verify the current worktree branch on the actual default-branch tip without using a default checkout.

how-agents-work

The cross-agent mental model for briefing many coding agents from one source: the doctrine layers, the one-capability-many-adapters principle, how a running agent picks up a new skill, and the boundary between personal source data and distributed artifacts.…

how-agents-work

The cross-agent mental model for briefing many coding agents from one source: the doctrine layers, the one-capability-many-adapters principle, how a running agent picks up a new skill, and the boundary between personal source data and distributed artifacts.…

naming-is-hard

Apply naming and wording doctrine to code symbols, files, branches, docs, UI copy, and mixed Dutch/English terms.

naming-is-hard

Apply naming and wording doctrine to code symbols, files, branches, docs, UI copy, and mixed Dutch/English terms.

parity-audit

The long invocation. A multi-agent, whole-codebase sweep over every entry point that promises one outcome, to find where their consequences diverge, converge the difference into the shared path, and re-run the audit to prove it moved. Runs as a named…

parity-audit

The long invocation. A multi-agent, whole-codebase sweep over every entry point that promises one outcome, to find where their consequences diverge, converge the difference into the shared path, and re-run the audit to prove it moved. Runs as a named…

rethink

Reconstruct a drifting project's purpose, boundaries, and promises from evidence, then write a governing manifesto before more local fixes.

rethink

Reconstruct a drifting project's purpose, boundaries, and promises from evidence, then write a governing manifesto before more local fixes.

drydry

The short invocation. Ask "have I seen this before?" about something you are looking at right now: duplicated logic, a parallel path, work that feels familiar while being written as new. Answers in seconds to minutes over a scope you name, and escalates to a…

drydry

The short invocation. Ask "have I seen this before?" about something you are looking at right now: duplicated logic, a parallel path, work that feels familiar while being written as new. Answers in seconds to minutes over a scope you name, and escalates to a…

second-opinion

Get a Claude second opinion on recent work or a design from Codex.

second-opinion

Get a Codex second opinion on recent work or a design from Claude Code.

bonsai

Create a ready git worktree from the freshest default branch and prime its `.bonsai` files. Worktree is sometimes misspelled as "work tree" (e.g. when dictated); treat both spellings as this skill.

prune

Safely remove a finished or abandoned git worktree while preserving unintegrated work.

bonsai

Create a ready git worktree from the freshest default branch and prime its `.bonsai` files. Worktree is sometimes misspelled as "work tree" (e.g. when dictated); treat both spellings as this skill.

prune

Safely remove a finished or abandoned git worktree while preserving unintegrated work.

saysay

Speech mode: speak every response aloud via macOS say. /saysay off to exit.

saysay

Speech mode: speak every response aloud via macOS say. /saysay off to exit.

art-director

Define visual identity and design-system foundations before UI work for new products, brand refreshes, or first design systems.

art-director

Define visual identity and design-system foundations before UI work for new products, brand refreshes, or first design systems.

programming-philosophy

Use when writing, refactoring, or reviewing code, especially state gates, root-cause fixes, DRY, and error handling.

programming-philosophy

Use when writing, refactoring, or reviewing code, especially state gates, root-cause fixes, DRY, and error handling.

pattern-memory

Search local Pattern Memory before implementing reusable code, interaction, or convention patterns.

pattern-memory

Search local Pattern Memory before implementing reusable code, interaction, or convention patterns.

not-your-monkey

Visual self-inspection correction for $not-your-monkey or frustration like "niet je aapje", "je aapje niet", "niet je bitch", or reload-aapje: inspect, fix, reload, and iterate instead of delegating visual QA.

not-your-monkey

Visual self-inspection correction for $not-your-monkey or frustration like "niet je aapje", "je aapje niet", "niet je bitch", or reload-aapje: inspect, fix, reload, and iterate instead of delegating visual QA.

verify

Define rover Done criteria and gather evidence against each criterion.

verify

Define rover Done criteria and gather evidence against each criterion.