skilo

github.com/yazcaleb/skilo

2026-03-06 ~ 2026-03-09 · 3 days

Perfectionist Paralysis

Aspirations for a universal AI skill sharing platform, suffocated under the weight of an endlessly polished landing page and premature trust mechanisms, all within 72 hours.

Perfected the welcome mat, forgot the house

Death Type

Perfectionist Paralysis

Skilo succumbed to 'Perfectionist Paralysis', fixated on an ideal state of presentation and architecture. 'packages/site/src/pages/Landing.tsx' alone saw over 2100 lines changed, while 'packages/api/src/utils/trust.ts' received 582 additions — an intricate trust mechanism for a system that never fully launched. The project was too busy perfecting its bow to ever launch the arrow.


Cause of Death

1. The endless landing page redesign

The 'packages/site/src/pages/Landing.tsx' file saw over 2100 lines changed (+1183 added, -950 removed) within 3 days. This suggests an intense focus on presentation, rather than core functionality, consuming significant development cycles on the project's visual facade.

2. Documentation over deployment

An estimated 30-40 of the 84 commits were dedicated to 'documentation and site polish', including 'Deslop README' and 'Polish docs, footer, and landing CTAs'. The 'packages/site/public/changelog.md' alone underwent 119 additions and 87 removals, indicating a meticulous, perhaps obsessive, approach to project self-description.

3. The 72-hour sprint to silence

The project accumulated 84 commits in a mere 3-day period (March 7-9, 2026), averaging 28 commits per day. This intense burst of activity, followed by abrupt silence, suggests a rapid acceleration into a wall of 'Perfectionist Paralysis', preventing sustained development beyond its initial, albeit frenetic, phase.


Vibe Score

27/ 100

AI-assisted but human-driven


What They Did

Skilo began with the grand ambition to democratize AI agent skills, allowing sharing and installation via simple links, targeting platforms like Claude Code, Codex, Cursor, and OpenCode. The project, a TypeScript monorepo using Turbo, aspired to provide a comprehensive solution: a 'skilo-cli' for command-line interaction and a 'skilo.xyz' website for management.

TypeScriptTurboReactInk

Burnout Analysis

The solo developer, 'Yaz', exhibited extreme dedication, landing 84 commits in 3 days, averaging 28 commits daily. A staggering 31% of this activity occurred late at night, and 82% on weekends. This short, intense burst of coding, followed by complete silence, is a textbook case of acute burnout, where ambition outpaced human endurance within 72 hours.


Dependency Archaeology

Despite aiming to support multiple AI platforms and offering both a web and CLI interface, the project surprisingly listed only 2 total dependencies. This suggests either an incredibly lean, custom implementation or, more likely, an anemic dependency tree at the moment of its demise, leaving its ambitious architecture unsupported by a robust external ecosystem.


Autopsy: File Structure

├──packages/site/src/pages/Landing.tsxOver 2100 lines changed: The Mona Lisa of unlaunched projects.
├──packages/api/src/utils/trust.tsOver 590 lines changed: Elaborate trust mechanisms for a system with no users to distrust.
├──packages/cli/src/ui/ink/selection.tsx245 lines added: A beautiful, interactive CLI UI for commands few would ever execute.
├──packages/site/public/changelog.md119 lines added, 87 removed: A meticulously documented history of a 3-day sprint.
├──packages/api/src/routes/skills.tsOver 680 lines changed: The complex heart of the skill sharing, beating strongly, but briefly.
├──packages/cli/src/commands/import.tsOver 430 lines changed: A robust import command, ready to ingest skills from an empty universe.
└──packages/api/src/utils/catalog.tsOver 410 lines changed: A comprehensive skill catalog, brimming with potential, yet empty.

Eulogy Stats

Total Commits
84
Ambitious Adjectives
0
Deploy Config
Yes
Estimated Users
0 (a perfectly polished welcome page for no one)

Last Words

The final commit messages, repeatedly 'Update generated changelog' (3 times), suggest the project's last breaths were spent meticulously cataloging its own, brief history, rather than forging a future.

May your next project's first commit be 'deploy to production', not 'polish footer'.

Perfectionist ParalysisAspirations for a universal AI skill sharing platform, suffocated under the weight of an endlessly polished landing page and premature trust mechanisms, all within 72 hours.

Share on X