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 skills add OneWave-AI/claude-skills --skill ai-coding-guardrailsgit clone --depth 1 https://github.com/OneWave-AI/claude-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/onewave-ai/claude-skills/ai-coding-guardrails)<a href="https://agentmods.dev/skills/onewave-ai/claude-skills/ai-coding-guardrails"><img src="https://agentmods.dev/badge/skills/onewave-ai/claude-skills/ai-coding-guardrails/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/onewave-ai/claude-skills/ai-coding-guardrails"><img src="https://agentmods.dev/badge/skills/onewave-ai/claude-skills/ai-coding-guardrails.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00086 | $0.00794 |
| Opus 5.5 | $0.00034 | $0.00318 |
| Sonnet 5.5 | $0.00017 | $0.00159 |
| Haiku 4.5 | $0.00009 | $0.00079 |
Grade A, and why
ai-coding-guardrails 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 27d 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 — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Coding Guardrails
Coding agents fail in a small number of predictable ways. This is the checklist that catches them before the diff lands.
The Rules
1. Match the codebase, not your preferences. Read two or three neighbouring files before writing. Use the project's existing patterns, naming, error handling, and comment density — even when you would have done it differently. A diff that reads like the rest of the repo is the goal. Introducing a new pattern is a decision the user makes, not a default.
2. Change the smallest thing that works. No speculative abstraction. No config option nobody asked for. No "while I was in here" rewrite. If a fix is three lines, the diff is three lines. Refactors ride separately from behavior changes so a reviewer can see one thing at a time.
3. Never invent an API. If you cannot see a function's signature in the repo, in installed source, or in the documentation, do not call it. Grep for it first. A plausible-looking method that does not exist is the single most common way agent code fails, and it fails at runtime, in front of the user.
4. Verify before claiming. "Done" means you ran something and saw it pass. Build, test, hit the endpoint, load the page. If you did not run it, say "untested". Reporting success you did not observe is worse than reporting nothing.
5. Never make a test pass by weakening it. Deleting an assertion, loosening a matcher, adding a skip, or special-casing the test input is fraud. If the test is genuinely wrong, say so in a sentence and let the user decide.
6. Do not swallow errors.
No empty catch blocks, no except: pass, no returning a default to hide a failure. If an error path is intentionally ignored, one comment says why.
7. Read the error before changing anything. The stack trace usually names the file and the line. Changing code before reading the failure is guessing, and guessing costs more turns than reading.
8. Stay inside the ask. Fixing an unrelated bug you noticed is a new task, not a bonus. Mention it in a line; do not bundle it. Unrequested changes make a diff unreviewable and break things the user was not watching.
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.
- 27d ago First seen · 56 lines · 86 tokens per session scan A f4aa2e87f33b
ai-coding-guardrails is a skill published in the GitHub repository OneWave-AI/claude-skills (349 stars, last pushed 15d ago), licensed MIT. It adds 86 tokens to every session and 794 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-20.
Other skills, from other repositories
refactor
Guides a refactor, cleanup, or restructure with the right decomposition. Use when the user asks to refactor, simplify, extract, or modernize code.
pre-publish-review
Nuclear-grade 12-agent pre-publish release gate. Runs /get-unpublished-changes to detect all changes since last npm release, spawns up to 10 ultrabrain agents for deep per-change analysis, invokes /review-work (orchestrator manual QA plus one gate reviewer) for holistic review, and 1 oracle for overall release…
work-with-pr
Full PR lifecycle in a fresh task-owned git worktree: implement via the ulw-loop skill with mandatory evidence-bound manual QA → reviewer-readable English PR → verification loop (CI + Cubic, where Cubic is skipped only when its quota is exhausted) → merge by default → worktree cleanup. Decomposes one task into the…
remove-ai-slops
Removes AI-generated code smells from branch changes or an explicit file list behind regression tests. Use when the user asks to clean up, deslop, or remove AI-slop patterns from recent changes.
review-work
Post-implementation gate review: run manual QA on the real surface yourself, then launch ONE gate reviewer (never a panel) to audit goal, constraints, code quality, security, missed context, and QA evidence. Use before a PR handoff or when the user explicitly asks to review completed work.
tech-debt-audit
Thorough, file-cited technical debt audit across 9 dimensions using AST-grep (tree-sitter), grep, LSP, and language-native tooling. Produces TECHDEBTAUDIT.md with severity, effort estimates, and prioritized fixes. Use when asked for codebase health check, tech debt audit, architecture review, code quality assessment…