Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/corichui/yolo/yolo-validategit clone --depth 1 https://github.com/CoriChui/yoloWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00035 | $0.01102 |
| Opus 5 | $0.00017 | $0.00551 |
| Sonnet 5 | $0.00007 | $0.00220 |
| Haiku 4.5 | $0.00003 | $0.00110 |
Grade A, and why
yolo-validate scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are running the YOLO change-validation loop documented in docs/validation/README.md.
Goal: determine whether the current state of the YOLO framework (skills + conventions +
templates + docs) is correct and genuinely helpful, surface concrete improvements, and —
when authorized — apply them and re-validate until green. This is a closed loop:
evaluate → synthesize → gate → act → iterate → report.
Arguments: $ARGUMENTS
- (none) → full evaluate + report, then STOP for confirmation before any fix.
--quick→ Layer 1 only (objective git harness). No agents, no fixes. Fast sanity check.--fix→ after presenting findings, apply the agreed fixes, then iterate.--auto→ no confirmation gate; apply high-confidence fixes and iterate autonomously, but STILL stop for anything touching landing/merge semantics, destructive changes, or low-confidence judgment calls.- a
target(branch name,diff, or path) → scope to that change. Default = working tree vsHEAD.
Cost & gating
Layer 1 is free and objective. Layers 2-4 spawn subagents (billed). Honor YOLO's own
cheap-vs-billed contract (.claude/yolo/conventions.md): evaluation runs, but STOP before
acting unless --fix/--auto. Beware correlated error — do not let one agent both
propose and bless a change; use the adversarial/swapped-order methods the suite prescribes.
Phase 1 — Evaluate (skip 2-4 if --quick)
- Layer 1 — objective, AI-free. Run
bash docs/validation/git-acceptance/run.sh. A red here is a hard failure; capture which case failed. No opinion required. - Layer 2 — adversarial red-team. Spawn a subagent with a hostile stance ("guilty until
proven innocent") over the target diff: find git/gh commands that won't run, internal
contradictions, config keys cited-but-undefined (cross-check
templates/config.yaml), and logic holes in the status derivation. It MUST reproduce findings with realgit/gh, not speculate. Pattern:docs/validation/adversarial/. - Layer 3 — behavioral. For each scenario in
docs/validation/behavioral/scenarios.md(plus any new decision point the target introduces), spawn a subagent that reads ONLY the relevantSKILL.md+conventions.mdand reports the decision the instructions yield; PASS/FAIL vs expected. Flag instructions ambiguous enough to be misread. - Layer 4 — blind A/B. For each user-facing doc changed in the target (
README.md,conventions.md,getting-started.md), extract OLD (git show HEAD:<file>) vs NEW (working tree), write both to neutral filenames in the scratchpad, and present them unlabeled as Variant A/B to two judge subagents with the A/B order swapped between them. Score on the rubric indocs/validation/ab-judge/rubric.md. Agreement across the swap is the signal; a split means investigate. Keep the answer key yourself; judges must not know which variant is the revision.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 71 lines · 35 tokens per session scan A 2d061f832640
yolo-validate is a command published in the GitHub repository CoriChui/yolo (9 stars, last pushed 1mo ago), licensed MIT. It adds 35 tokens to every session and 1,102 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
sm-sail
Command "sm-sail" from ScienceIsNeato/slop-mop, covering /sm-sail — drive a pr to green, autonomously, the loop, when sail parks on review threads, when to stop before "pr ready" — only two reasons and expect convergence, not one pass.
review
Cold re-quiz on code that already shipped — your own session commits, not the change in front of you.
sm-barnacle
Use when sm itself gives invalid guidance, blocks valid work, produces confusing output, or breaks install/upgrade/refit flow. Do not use this for real target-repo failures; fix those through the normal rail.
sm-scour
Run slop-mop's comprehensive pre-PR sweep for this repository.
content-ideas
Build today's For You feed — scrape tracked competitors across X, Instagram, TikTok, and YouTube, score what's performing, and turn it into actionable content ideas.
review-pr
PR을 설계문서·이슈 계약 대비로 검증 후 승인/반려 판정.