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 TheGoat395/Codex-Skills --skill code-change-safety-checkpointgit clone --depth 1 https://github.com/TheGoat395/Codex-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/thegoat395/codex-skills/code-change-safety-checkpoint)<a href="https://agentmods.dev/skills/thegoat395/codex-skills/code-change-safety-checkpoint"><img src="https://agentmods.dev/badge/skills/thegoat395/codex-skills/code-change-safety-checkpoint/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/thegoat395/codex-skills/code-change-safety-checkpoint"><img src="https://agentmods.dev/badge/skills/thegoat395/codex-skills/code-change-safety-checkpoint.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Tool Misuse · line 59 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
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 | $0.00015 | $0.00538 |
| Opus 5 | $0.00008 | $0.00269 |
| Sonnet 5 | $0.00003 | $0.00108 |
| Haiku 4.5 | $0.00002 | $0.00054 |
Grade A, and why
code-change-safety-checkpoint 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 4d 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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Code Change Safety Checkpoint
Use this skill before risky edits. The goal is simple: make sure there is a credible way back before changing files.
Trigger
Use when:
- the user asks for website, app, repo, or code changes with nontrivial blast radius
- files may be deleted, moved, regenerated, overwritten, or mass-edited
- the repo has uncommitted changes
- the work touches shared behavior, routing, deployment config, generated assets, or build tooling
- the user specifically asks for backup, checkpoint, rollback protection, or protection from accidental destructive changes
Do not run heavy backup steps for tiny read-only analysis or single-line edits unless the user asks.
Workflow
-
Inspect state before editing.
- Run
git status --short --branchwhen inside a git repo. - Identify changed files that already exist before your work.
- Treat uncommitted user changes as protected.
- Run
-
Choose the rollback path.
- Follow the current repository/worktree branch policy. Record a recoverable base commit for tracked data; a branch label alone does not preserve uncommitted or untracked files.
- If the repo has user changes, preserve the affected hunks and untracked targets with an appropriate patch, snapshot or scoped backup when the edit could destroy them. Avoid redundant copies of already protected targets.
- If there is no git repo or the work involves generated assets, make a timestamped local backup of the files or folders you will touch.
-
Explain planned edits before changing files.
- List the main files or directories you expect to edit.
- Call out any deletion, move, or overwrite.
- A clear task instruction authorizes routine in-scope edits regardless of its verbs. Ask only for a materially ambiguous destructive target or another genuinely missing authorization.
-
Edit narrowly.
- Avoid unrelated refactors.
- Do not revert user changes unless explicitly requested.
- Prefer additive changes when the risk is unclear.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago Changed · -39 tokens per session da8ebbc8dc0e
- 10d ago First seen · 62 lines · 54 tokens per session scan A 257de41d7824
code-change-safety-checkpoint is a skill published in the GitHub repository TheGoat395/Codex-Skills (122 stars, last pushed yesterday), licensed MIT. It adds 15 tokens to every session and 538 once invoked, about $0.0001 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 2026-08-30.
Other skills, from other repositories
image-to-threejs
REQUIRES uv + Python 3.12 and a one-time git clone — unlike the other asset skills, this one is not Node-only. Turn a generated reference image into a rigged, procedural Three.js prop — code, not a mesh file. Generate ONE clean product-shot reference with vg generate, author a component spec (parts, materials, pivots…
game-ui
Game interface craft — HUD zoning and hierarchy, fixed-width numerals, alert-color roles, meters over stat cards, menu/overlay states, touch controls (44px targets, safe-area insets, pointer-release handling) — for Phaser, Three.js, and DOM overlays. Use when: building or reviewing a HUD, score/timer/health display…
design-pipeline
A build-and-test workflow for turning Figma designs into shadcn/ui components, a reusable set of interface components for React-based projects. It checks the result in a real browser at three screen sizes and runs accessibility checks.
swiftui-animation
Implement, diagnose, or review SwiftUI motion using explicit and scoped implicit animations, springs, transitions, PhaseAnimator, KeyframeAnimator, matched geometry or navigation zoom, SF Symbol effects, and custom Animation types. Use when views should animate on state changes, insertion, removal, navigation, or…
swiftui-gestures
Implement, review, or improve SwiftUI gesture handling. Use when adding tap, long press, drag, magnify, or rotate gestures, composing gestures with simultaneously/sequenced/exclusively, managing transient state with @GestureState, resolving parent/child gesture conflicts with highPriorityGesture or…
focus-engine
Implements keyboard, directional, and scene-level focus behavior across SwiftUI and UIKit. Use when managing @FocusState, defaultFocus, focused values, focusable interactions, focus sections, tvOS geometric focus and Siri Remote navigation, watchOS Digital Crown input, visionOS connected-device focus versus gaze…