An agent skill that turns "how could this project make money?" into a scanned inventory, sourced market research, a gap analysis, and action specs an agent can execute.
Most monetization advice fails in one of two ways: it's generated from the model's memory of your project instead of the project itself, or it ends at "you could build a SaaS" — a recommendation nobody can act on. This skill forbids both.
Every claim traces to either a file in the repo or a dated source on the internet. Analysis that cites neither is opinion and gets cut.
Four phases, each with an enforced output contract:
- Scan the project, not the pitch — inventory sellable surfaces from the repo and running state (dead code and flag-gated demos don't count; consumers and routes are verified), existing monetization infrastructure (its absence is an explicit finding, and it usually is absent), distribution state, and the compliance artifacts the target market will demand.
- Research with receipts — comparable products and their actual pricing, the category's dominant monetization pattern, regulatory or procurement forcing functions that compel buying, and distribution channels. Every finding carries a URL and access date; prior knowledge without a source is a hypothesis to verify, not a finding.
- Gap analysis — diff the two in both directions: opportunity gaps (the market pays for X, the project has X, nothing connects them) and execution gaps (what's missing to actually charge — billing, tiers, license, compliance artifact, case study). Ranked by distance-to-revenue, not by size of opportunity.
- Action specs, not recommendations — every proposal is a spec with numbered agent steps naming files and commands, evidence citations (minimum two independent sources for anything that gates real work), measurable verification, and a mandatory human-only list: account creation, payment setup, legal signoff, and publish decisions are never silently absorbed by the agent.
Two kinds of spec are legal: product specs (the deliverable is sold) and enabler specs (the deliverable unblocks a sale — a conformance report, a case study, a pricing page — with the revenue hypothesis naming the downstream sale).
This skill was written against a documented baseline failure (an audit that scanned nothing, researched nothing, and produced advice), then tested by having a fresh agent execute it against a real production codebase. The test agent's six ambiguity reports were each resolved and re-verified. The output-format rules in the skill are the ones that survived that loop.
Pairs with skill-productizer's 13-forms taxonomy (the vocabulary for Phase 4's Form: line) and productize-real-assets (cross-project personal asset inventory). This skill is the per-project deep audit.
npx skills add succtorlin/monetization-audit(installs for Claude Code, Codex, Cursor, Gemini CLI and other agents via the skills CLI) — or manually:
git clone https://github.com/succtorlin/monetization-audit.git
cp -r monetization-audit/skills/monetization-audit ~/.claude/skills/Once installed, Claude picks it up on prompts like:
- "How can this project make money?"
- "What monetization opportunities are we missing?"
- "Research pricing and competitors for this product"
- "Turn this codebase into a revenue plan I can act on"
Bounded runs are supported — "just run the scan phase" produces the Phase 1 inventory with the scope stated, no full report required.
A dated report at docs/monetization/<YYYY-MM-DD>-audit.md in the audited project, containing all four phase outputs. The report is the deliverable; chat is just the summary.
MIT — see LICENSE.