Skip to main content

ADR-013: Distribution Channel Priority

Date: 2026-06-24
Status: Accepted

Context

Agenthood has four distribution channels available simultaneously. With a single maintainer and limited time, investing equally in all four is a mistake. This ADR decides which channel gets optimised first, second, and third — and what the criteria are for re-prioritising.

The four channels:

  1. npm — npm install --save-dev agenthood. Published automatically via @semantic-release/npm on every main push. Currently published as latest (managed by semantic-release). Requires keyword optimisation for discoverability.
  2. Skills.sh — 600k+ skills marketplace, Vercel-backed. Skills appear automatically once installed via npx agenthood init. Zero-cost listing via skills.sh.json.
  3. VS Code Marketplace — IDE extension (agenthood-vscode). Published via .github/workflows/vscode-extension.yml. Currently at v0.1.0. Requires publisher account setup and ongoing maintenance.
  4. SkillsMP — 1.7M+ skills, GitHub-scrape-based. Requires 2+ stars for listing. Star count currently insufficient.

Decision

Priority order (highest to lowest):

  1. npm — Already published, keyword optimisation is immediate. The primary adoption trigger for the CLI. One package.json change + compounding discoverability via weekly download stats.
  2. Skills.sh — Zero additional cost. Once npx agenthood init is run, skills appear automatically on the leaderboard. Seed installs are the only activation needed. Groupings already configured in skills.sh.json.
  3. VS Code Marketplace — Requires publisher account verification (1–2 hours setup). High ongoing visibility but lower engagement intensity than CLI users. Extension at v0.1.0 — feature-complete for web views but not yet a primary adoption driver.
  4. SkillsMP — Blocked by the 2-star gate. Comes naturally as npm + skills.sh drive star growth.

This order reflects the minimum effort for maximum incremental visibility principle. Each channel is optimised only when the previous channel is fully operational.

Alternatives Considered

OptionProsConsWhy Rejected
Skills.sh firstHighest traffic (600k+), Vercel-backed leaderboardLeaderboard position is purely install-count; gaming is possibleWithout npm adoption, there are no seed installs to drive the leaderboard
SkillsMP first1.7M+ skills, GitHub-scrape-based (no manual submission)Requires 2+ stars gate; no control over listing timingBlocked by current star count — cannot act on it now
VS Code Marketplace firstDirect IDE integration; one-click trialLower engagement than CLI users; publisher setup overheadNot yet feature-complete enough to justify the setup investment

Consequences

Easier:

  • Clear ordering removes decision paralysis — always know which channel to work on next
  • npm and Skills.sh compound benefits: npm downloads increase stars, stars unlock SkillsMP
  • Semantic release pipeline already publishes npm automatically — no new infrastructure needed

Harder:

  • VS Code Marketplace and SkillsMP remain inactive until npm and Skills.sh reach targets
  • If npm keyword optimisation does not move the needle within 4 weeks, the entire strategy pivots

Re-prioritisation trigger:
If npm weekly downloads do not reach 50 within 4 weeks of keyword optimisation, escalate VS Code Marketplace to #2 priority and begin active publisher setup. Skills.sh and SkillsMP priorities remain unchanged in this scenario — Skills.sh is already zero-cost and SkillsMP remains star-gated regardless of channel priority.

References