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 skills/mghareeb/code-change-impact/code-change-impactnpx skills add mghareeb/code-change-impact --skill code-change-impactgit clone --depth 1 https://github.com/mghareeb/code-change-impactWhat 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 | $0.00238 | $0.02586 |
| Opus 5 | $0.00119 | $0.01293 |
| Sonnet 5 | $0.00048 | $0.00517 |
| Haiku 4.5 | $0.00024 | $0.00259 |
Grade A, and why
code-change-impact 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 yesterday.
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 — 213 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Code Change Impact — blast-radius / regression analysis for any codebase
A fix is "done" only when you know what it touched besides the thing you were fixing. The dangerous ripples are rarely in the file you edited — they're in the callers of a function whose behavior you changed, the other modules that import a shared helper, the client that decodes a response whose shape you altered, the generated/duplicated twin file you forgot to regenerate, or a consumer that now fails to compile. This skill finds those, then proves the answer by running the project's own build/test/lint and exercising the impacted surfaces.
Run it after a change, against the VCS diff (the diff is the source of truth for "what changed"). It adapts to the language and tooling of whatever repo it's in — so Phase 0 (discovery) comes first and the rest builds on it.
Phase 0 — Discover the project (do this before analyzing)
You can't trace dependencies or verify a build until you know the project's
conventions. Spend a minute learning them; cache the findings for the rest of the
run. See references/recipes.md §1–2 for per-ecosystem detection commands.
Establish:
- Repo root + diff —
git rev-parse --show-toplevel; then the change set (working tree, staged, or branch-vs-base). Use the VCS in play (git/hg/jj). - Languages + ecosystems — from manifests and file extensions
(
package.json,pyproject.toml/requirements.txt,go.mod,pom.xml/build.gradle,*.csproj/*.sln,Gemfile,Cargo.toml,composer.json). A repo may have several (e.g. a JS frontend + a Go backend). - Verification commands — the project's own typecheck, build, test, and lint
commands. Read them from the manifest's script section / Makefile / CI config
rather than guessing (
references/recipes.md§2). Note which test scripts are targeted (named per area) — those let you verify a specific blast radius cheaply. - Module + import style — how this language names and imports modules, and
any path aliases (e.g. a TS
tsconfig.jsonpaths, a Python namespace, a Go module path). This is what your reverse-dependency greps will key on.
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.
- yesterday First seen · 213 lines · 238 tokens per session scan A b3235f524874
code-change-impact is a skill published in the GitHub repository mghareeb/code-change-impact (2 stars, last pushed 2mo ago), licensed MIT. It adds 238 tokens to every session and 2,586 once invoked, about $0.0012 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 skills, from other repositories
karpathy-llm-wiki
Use when building or maintaining a personal LLM-powered knowledge base. Triggers: ingesting sources into a wiki, querying wiki knowledge, linting wiki quality, 'add to wiki', 'what do I know about', or any mention of 'LLM wiki' or 'Karpathy wiki'.
nopua
The anti-PUA. Drives AI with wisdom, trust, and inner motivation instead of fear and threats. Activates on: task failed 2+ times, about to give up, suggesting user do it manually, blaming environment unverified, stuck in loops, passive behavior, or user frustration ('try harder', 'figure it out', '换个方法', '为什么还不行').…
nopua-lite
NoPUA Lite — core wisdom in 1.5k tokens. Drives AI with trust and inner motivation instead of fear. Same Daoist philosophy, minimal footprint. For personal use and small-context models.
resemble-detect
Deepfake detection and media safety — detect AI-generated audio, images, and video, trace synthesis sources, and analyze media intelligence using direct Resemble AI API calls.
ios-agent-skill
Expert iOS/Swift developer behavior for AI coding agents. Use when writing, reviewing, or refactoring Swift, SwiftUI, UIKit, or SwiftData code; when designing iOS app architecture (MVVM, Clean Architecture, coordinators, routing); when building UI that must meet Apple's Human Interface Guidelines, contrast, dark-mode…
cinematic-scroll
Build cinematic scroll-driven, 3D-tilt, parallax, and environment-morphing websites — pinned chapter reveals, hero parallax, depth-image figures, hover-tilt cards, background-morphing layouts, release/launch pages, product story pages, or editorial commerce microsites. From a single self-contained scroll section (Mode…