Andrew Luxem

Open source

Code and systems, in public

Agent skills, design tooling, and demo pipelines from the practice. Nothing here is a slide about AI. It is public code and provenance, reviewed before adoption.

16 REPOS · 14 AGENT SKILLS · 8-CHECK REVIEW · github.com/andrewluxem

Agent skills

Portable skills, no external network access.

customer-journey-lifecycle

A portable agent skill for customer-state modeling, lifecycle journey design, contact governance, and incremental measurement.

SKILL.MD · LIFECYCLE · MEASUREMENT

View on GitHub
generational-marketing-review

A portable agent skill for reviewing marketing plans and assets for generational audience fit.

SKILL.MD · AUDIENCE · REVIEW

View on GitHub
decision-card

A static Claude skill that stress-tests a decision and returns an answer-first card. Explicit invocation only: no self-firing, no remote fetch, no auto-update.

SKILL.MD · DECISIONS · STATIC

View on GitHub
successful-meetings

A portable agent skill for meeting prep, follow-up notes, and meeting audits.

SKILL.MD · MEETINGS · STATIC

View on GitHub
3ps-framework

A portable agent skill that turns a week of raw notes into one weekly team update under three headings: Progress, Plans, Problems.

SKILL.MD · STATUS UPDATES · STATIC

View on GitHub
silent-meetings

A portable agent skill for read-first meetings: pre-read documents, meeting packets, and readiness checks.

SKILL.MD · MEETINGS · STATIC

View on GitHub
4-blocker-business-reviews

A portable agent skill for the weekly business review: four blocks, goals written as controllable inputs, and a draft check with a verdict.

SKILL.MD · BUSINESS REVIEW · STATIC

View on GitHub
weekly-status-updates

A portable agent skill for weekly project status updates: position against the plan, a status call that checks dependencies, and a draft check with a verdict.

SKILL.MD · STATUS UPDATES · STATIC

View on GitHub
weekly-schedule-of-meetings

A portable agent skill for designing a team week: a weekly cadence with themed days and input deadlines, an operating calendar for what does not run weekly, and an audit of a week that already exists.

SKILL.MD · MEETING CADENCE · STATIC

View on GitHub
10x-strategy-meeting

A portable agent skill for running a 10x strategy session: a session packet with groups, roles, and starter questions tailored to the business, and a capture that routes every idea to an owner and a date.

SKILL.MD · STRATEGY · STATIC

View on GitHub
prioritization-formula

A portable agent skill that scores and ranks competing work with the arithmetic shown: a scored backlog where every row derives its own priority and every unsupported value carries a labeled assumption, plus a sniff test that challenges a submitted ranking without rewriting its scores.

SKILL.MD · PLANNING · STATIC

View on GitHub
business-goals

A portable agent skill that turns stated intentions into a measurable goals document built on the ten questions, where a goal with no supplied number reads Target needed and a question the input never answered stays in the page as a labeled gap, plus a checkpoint review that calls each goal against reported actuals.

SKILL.MD · PLANNING · STATIC

View on GitHub
loglines

A portable agent skill that plans a launch by writing it as already shipped, a future-dated logline carrying one ranked customer benefit, customer and stakeholder FAQs where absent research reads Evidence needed and an unsourced job story reads Source needed, and a mandatory out-of-scope, plus a 13-check critique of an existing draft that ends in a verdict.

SKILL.MD · PLANNING · STATIC

View on GitHub
how-to-plan-new-product-and-services

A portable agent skill that turns an idea brief into a six-phase product plan, where alpha to beta and beta to GA are explicit go or no-go rows, the instrumentation row lands before the beta widens, an unassigned step reads Owner needed and an unsupplied date runs as an offset rather than a guess, plus a bet review that buckets competing initiatives and checks allocation against the 70 20 10 guide.

SKILL.MD · PLANNING · STATIC

View on GitHub

Method

The audit before adoption

Every repository, fork, and third-party tool goes through the same eight-point review before it is adopted or recommended: where it came from, how it is licensed, what it depends on, whether it reaches the network, how it is invoked, how it updates, how it handles data, and how much it can break. The skepticism I bring to a vendor deck, pointed at code.

  • 01 PROVENANCE
  • 02 LICENSE
  • 03 DEPENDENCIES
  • 04 NETWORK CALLS
  • 05 INVOCATION
  • 06 UPDATE PATH
  • 07 DATA HANDLING
  • 08 BLAST RADIUS

The playbook library is built for execution

The library is structured for execution. All 51 free playbooks currently ship as downloadable skills. Premium playbooks remain visible in the library and do not receive an Install tab until their source packages are complete.

Browse the skills library