Zero Tech Debt
Zero Tech Debt is a compact agent skill for reworking a recent change as if the intended UX and architecture existed from day one, deleting compatibility cruft and accidental complexity instead of preserving the scar tissue of how the change was made. Source: X/@iamsahaj_xyz, 2026-05-14
What Is Verifiable
The deep review found the missing artifact: the self-thread/artifact reply contains the complete SKILL.md body, and the local image artifact matches it. There is still no verified registry slug or repo release, so this should be treated as a source-captured custom skill rather than a skills.sh package. Source: raw X bookmark enrichment and local visual review, 2026-07-01
The actual skill is not a generic tech-debt scanner. Its specific niche is post-change shape cleanup:
- state the intended end state
- search real callers before preserving compatibility
- delete unused modes, props, wrappers, route aliases, fallbacks, and shims
- reshape around the final product surface
- move shared rules to one owner
- verify the intended flow and deleted assumptions
Kevin Stack Decision
This is now installed as skills/engineering/zero-tech-debt/SKILL.md because it has a distinct trigger from Security and Review Skills, Security and Review Skills, and Frontend and Design Skills. Those skills review, improve, or plan. zero-tech-debt is a focused cleanup pass after a feature lands in a patched-on shape.
The skill should be used when a change works but carries accidental complexity from the implementation path: leftover compatibility branches, mode flags, wrapper components, defensive fallbacks, duplicated permission gates, or names that describe history instead of product intent.
Evaluation
- Use after implementation. It is strongest after an AI-generated or hurried change works but feels like layers of preserved history.
- Do not use as a broad audit. For broad read-only architecture review, use Frontend and Design Skills or Security and Review Skills.
- Do not make it a framework generator. The rule is to simplify toward the real product surface, not invent a generic abstraction.
- Security remains separate. Run Security and Review Skills or security-doctor-chain when the cleanup touches auth, data, secrets, permissions, or external calls.
Timeline
- 2026-05-14 | Announced on X by @iamsahaj_xyz. 970 likes, 1,447 bookmarks. Source: X/@iamsahaj_xyz, 2026-05-14
- 2026-06-13 | Quality remediation: searched GitHub/web for a canonical artifact — none attributable to @iamsahaj_xyz found; reframed as thin candidate, mapped overlap to owned skills (Security and Review Skills/Security and Review Skills/Security and Review Skills). Source: https://x.com/iamsahaj_xyz/status/2054998319232065707
- 2026-07-01 | Deep review found the full skill body in the X self-thread/artifact reply and local image artifact; installed it as a focused engineering skill for deleting compatibility cruft after a recent change. Source: raw/x-bookmarks/enriched/2054998319232065707.json; local visual review