Ponytail: vendor claims and pricing as captured 2026-10-29 ~08:50-08:56 UTC+8 (checked_on 2026-10-29)

Sources
- README.md at repo HEAD 6309875aa27a788030b8e3f43e29a1342a6370ee (commit 2026-10-29 08:00 UTC+8): https://github.com/DietrichGebert/ponytail
- benchmarks/README.md (same commit)
- https://ponytail.dev/ (HTTP 200)
- npm @dietrichgebert/ponytail 5.3.2 (latest at check time)
- GitHub API: 179,198 stars, MIT, last push 2026-10-29 08:00 UTC+8

Pricing: free, MIT license, no account. README links a waitlist (ponytail.dev/soon) and merch; no paid tier for the skill seen.

Verbatim claims
1. README hero: "Ponytail 5: rebuilt from the ground up. -53% code · -41% time · -26% cost · -45% tokens. And yet: 98% of risky logic ships with a test. Without Ponytail: 68%." ... "Benchmarked in Claude Code, the same agent with and without the skill: 39 tasks including a real FastAPI + React repo, Opus 5.5, 5 runs each."
2. README "How it works": "Lazy, not negligent: trust-boundary validation, data-loss handling, security, and accessibility are never on the chopping block."
3. README "How it works": "Logic with a branch, a loop, a parser, money or security leaves one small test behind. Every reply ends with what was skipped or not checked and any risk you should know."
4. README "The prompt": "Ponytail is one prompt: skills/ponytail/SKILL.md. The compact version, for agents that read a rules file, is AGENTS.md. Everything else in this repo loads that prompt into different agents."
5. README Install: "Any other agent: copy AGENTS.md into your project, or ask your agent to install skills/ponytail/SKILL.md as a skill."
6. ponytail.dev: "A ruleset that makes your AI coding agent write the least code that works: 53% less code, 41% faster, 26% cheaper, and just as correct." ... "Validation, error handling, security and accessibility are never simplified away."
7. benchmarks/README.md: single-shot benchmark arm = SKILL.md as the system prompt (benchmarks/arms/ponytail.js: "the repo's own SKILL.md (full) as the system prompt").