AgentLead's Skills · Claude Code · Codex · Antigravity · OpenCode
AgentLead's Skills encode engineering discipline so Claude follows it by default. You don't choose a skill — you describe the task and the right flow fires: investigate before building, design the interface first, test-first, review at the boundary, verify before merge.
Match your situation to an entry point. The flow takes it from there.
feature-doc
Standard feature →
Large feature — many ACs or multi-package
feature-doc + tdd-rounds
Large feature →
New feature, direction still unclear
investigate
Investigation →
Bug — root cause obvious from the trace
tdd
Bug fix →
Bug — intermittent, flaky, or cause unclear
debug → tdd
Debug flow →
Refactor / improve existing architecture
improve-codebase-architecture
Refactor →
New multi-module system from scratch
system-design
Greenfield system →
Reviewing someone else's PR
pr-review
Review →
Each is a chain of skills. Doc steps write a contract; gate steps refuse to advance until they pass.
A single-package feature with a manageable acceptance-criteria count.
Orchestrated across rounds — the parent dispatches Builder sub-agents and holds state.
For new dependency, framework choice, or a cross-cutting refactor — options before code.
The symptom and stack trace point straight at the bug.
Environment-specific, a regression, or you don't yet know what to assert.
Improve architecture — turn shallow modules into deep ones. One candidate at a time.
A new multi-module project. After topology is set, each feature runs its own flow.
Runs whenever you read a diff — classifies findings by severity.
These don't have a flow of their own — they hook into the ones above.
Boring code, naming as the primary refactor, YAGNI, rule of 3, locality. Carried into simplify and pr-review.
Fires on surface-changing work — new entry points, identity flows, authz, sensitive data, external deps. Runs alongside tdd.
Interrupts any flow when you're lost in unfamiliar code. Maps the relevant modules and callers, then hands back.
Performance benchmarking discipline — measures latency and throughput when a flow carries a performance AC.
One command. Works for Claude Code, Codex, Antigravity, and OpenCode.
# interactive — pick assistants & scope
$ npx @ai-agent-lead/skills
# everything, no prompts
$ npx @ai-agent-lead/skills --all --force
Scope it with --global (~/.claude/skills) or --local
(./.claude/skills); target one assistant with --claude,
--codex, --antigravity, or --opencode.