patch-tracker

patch-tracker is a skill for Claude Code, Codex from EZotoff/ez-omo-config. It costs 33 tokens per session (4,675 once invoked), scanned A, original, MIT.

A registry and checking process for custom changes made directly to external dependencies. External dependencies are libraries or packages maintained outside your project.

In plain words
What is it for?
Use it to register, audit and verify dependency patches, especially fixes that may be lost during upgrades.
Why use it?
It helps record those changes and check whether they still work after a dependency update.

Skill for Claude CodeCodex

Install

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.

agentmods
npx agentmods add skills/ezotoff/ez-omo-config/patch-tracker
Any agent
npx skills add EZotoff/ez-omo-config --skill patch-tracker
Clone the repo
git clone --depth 1 https://github.com/EZotoff/ez-omo-config

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for patch-tracker

README.md
[![agentmods](https://agentmods.dev/badge/skills/ezotoff/ez-omo-config/patch-tracker.svg)](https://agentmods.dev/skills/ezotoff/ez-omo-config/patch-tracker)
Your own site
<a href="https://agentmods.dev/skills/ezotoff/ez-omo-config/patch-tracker"><img src="https://agentmods.dev/badge/skills/ezotoff/ez-omo-config/patch-tracker.svg" alt="Measured on agentmods" height="20"></a>
Per session 33 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,675 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00033 $0.04675
Opus 5 $0.00016 $0.02337
Sonnet 5 $0.00007 $0.00935
Haiku 4.5 $0.00003 $0.00468

Measured yesterday against content hash 681c81d4be53, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

patch-tracker 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.

skills/patch-tracker/SKILL.md · 376 lines

How it starts

The opening of the file, as written. The whole thing — 376 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Patch Tracker — Skill for Dependency Patch CRUD and Verification

Triggers: /patch-tracker, /check-patches, /register-patch, patch external, custom fix, won't survive update, dependency hack, check patches, verify patches, track patches, patch registry, dependency modifications


⚠️ DURABLE ALTERNATIVE NUDGE — READ BEFORE PATCHING

BEFORE creating a direct patch, ALWAYS evaluate these alternatives in order:

  1. Can this be done via a plugin/hook? (PREFERRED — survives updates automatically)
  2. Can this be done via configuration? (PREFERRED — portable, no code changes)
  3. Can this be upstreamed? (PREFERRED — benefits everyone, eliminates maintenance)
  4. Is a direct patch the ONLY option? (LAST RESORT — register it below)

Direct patches are technical debt by design. Every patch registered here is a liability that must be re-verified after every dependency update. If a durable alternative exists, use it instead.


WORKFLOW

1. CREATE — Register a New Patch

When invoked with /register-patch or when a custom patch is applied:

Step 1: Evaluate durable alternatives (see nudge above). If a plugin, hook, or config option exists, use that instead and do NOT create a patch entry.

Step 2: Gather required frontmatter fields (see .sisyphus/patches/TEMPLATE.md):

Field Required Description
patch_id yes Unique slug: {dep}--{short-description} (e.g., omo--notepad-directive-scoping)
dependency yes Package or tool name being patched
target_file yes Relative path within the dependency install directory
target_install_path yes Absolute path to the dependency install root
status yes active | upstreamed | deprecated
applied_date yes ISO 8601 date when patch was applied (YYYY-MM-DD)
dep_version yes Version or range where the patch was LAST VERIFIED EFFECTIVE at runtime. Do NOT bump this just because the patch string is present in a newer binary — that is a pattern-presence claim, not an effectiveness claim (see runtime_effective).
upstream_issue yes URL to upstream issue/PR tracking the fix, or "none"
verification_pattern yes Grep-compatible regex to verify the patch string is embedded. Limitation: Bun minification preserves JS property keys and string literals, so a property-key pattern (e.g. __linkLabelPatch) reports a false-positive APPLIED even when the surrounding code is unreachable. Prefer a pattern that disappears if the code path is dead; otherwise require a ## Runtime Verification section.
surfaces rendering patches only List of surfaces the patch touches (cli-run, tui-interactive, server-api). REQUIRED for any patch whose target_file is in a rendering directory (cli/cmd/run/, tui/src/routes/, etc.) or that overrides a renderable method.
runtime_effective monkey-patches only Boolean. REQUIRED for any patch that overrides an internal method, attaches a ref callback, or patches a renderable. Set true only after observing the patched behaviour on the real surface. Set false when the patch string is present but the feature regressed (see opencode--link-click-wrapped-osc8.md).
upstream_issue yes URL to upstream issue/PR tracking the fix, or "none"
verification_pattern yes Grep-compatible regex to verify patch is applied

Read the full file on GitHub · 376 lines

Changes

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.

  1. yesterday Changed · +18 lines 681c81d4be53
  2. 5d ago First seen · 358 lines · 33 tokens per session scan A 0a7c3d89f2a3

Subscribe to this mod's changes

patch-tracker is a skill published in the GitHub repository EZotoff/ez-omo-config (5 stars, last pushed 4d ago), licensed MIT. It adds 33 tokens to every session and 4,675 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.