Fresh Data

📡 Always fetch fresh data when relevant to your agentic development. 📡

The Problem

AI models can default to their knowledge relative to it's training data. This can be for key information related to problem solving which can draw to the wrong analysis or problems. This skill, or AI-model harness extension aims to enforce the process of fetching the most up-to-date, revelant data available when applicable.

How it works

Before citing key facts, the agent stops at the first rung that holds:

1. Does this claim age?         → no: cite from training (math, history, ideas)
2. What time is it?             → `date` for knowing the "retrieved today"
3. Primary source exists?       → fetch it (vendor docs, changelogs, registries)
4. Is the URL authoritative?    → state it; unfamiliar → click through to vendor homepage
5. If passes:                   → quote with URL and retrieval timestamp

Otherwise, if sources aren't authoritative or can't be fetched fresh data will state this that information couldn't be found.

When it doesn't apply

Fresh data will ignore claims where training-era data is fine:

• Historical events with fixed dates — e.g. "When did WWII end?"
• Mathematical or logical facts — pure math, formal logic, well-established identities
• Foundational concepts that do not evolve — e.g. what TCP/IP is, what a hash function does
• Blog posts, opinion pieces, essays — when you want the idea or argument, not a specific factual claim that the post hinges on
• Discovering new opinions or perspectives — when the recency of the view doesn't matter; historic articles are fine

Key examples

• Using debian-12 instead of debian-13.
• Fetching previous version of dependencies.
• Quoting an API endpoint URL that's since been deprecated or moved.
• Citing an old default version of a framework (Next.js, React, Python, Node).
• Naming a CVE or security advisory that has since been patched or superseded.
• Using a deprecated CLI flag (e.g. old kubectl / docker / gh syntax).

Worked examples

Check the latest version of an npm package:

curl -s https://registry.npmjs.org/<pkg>/latest | jq .version
# Or use the package manager:
npm view <pkg> version

Check peer / engine compatibility without guessing:

npm view <pkg> peerDependencies engines

Find the newest articles on a site:

curl -s https://<site>/sitemap.xml | grep -E '<loc>|<lastmod>'
# Sort by <lastmod> descending. For large sites, try /sitemap_index.xml first —
# it links to per-section sitemaps via <sitemap><loc> entries.

Read a feed directly:

curl -s https://<site>/rss.xml | grep -E '<title>|<pubDate>'

These are the canonical primitives. See SKILL.md for the full procedure and the URL list for PyPI / crates.io / RubyGems / Maven / Packagist / NuGet / Go / container registries.

Install

Pi

pi install npm:@mcwarlus/fresh-data
# or pin to a specific git ref:
pi install git:mcwarlus/fresh-data@<ref>

Once published, npm is the canonical install path. Git install remains available for tag pinning.

Once installed:

• Slash commands:
  • /fresh-data — toggle the rules on/off
  • /fresh-data on or /fresh-data enable — explicitly turn the rules on
  • /fresh-data off or /fresh-data disable — explicitly turn the rules off
  • /fresh-data status — report current state
  • /fresh-data install — write the rules into CLAUDE.md so they apply even when the skill is off
• Verbal toggle: type stop fresh-data or normal mode to turn the rules off
• Footer indicator: ● 📡 fresh-data: ACTIVE vs ○ 📡 fresh-data: OFF
• Default: fresh sessions start on after install; turning it off in a session persists across /resume and /new.

Claude Code / skills.sh

npx skills@latest add mcwarlus/fresh-data

Clone-based harnesses

Clone the repo; each harness auto-loads its file from the path below:

Harness	Path it loads
Cursor	.cursor/rules/fresh-data.mdc
Codex	.codex-plugin/plugin.json
Cline	.clinerules/fresh-data.md
Qoder	.qoder/rules/fresh-data.md
Kiro	.kiro/steering/fresh-data.md
OpenClaw	.openclaw/skills/fresh-data/SKILL.md
OpenCode	.opencode/command/fresh-data.md

Development

npm test                       # runs all 57 tests (~190ms)
node scripts/check-rule-copies.js

Contributing

Issues and PRs at github.com/mcwarlus/fresh-data.

The rules body is canonical in RULES.md; every harness-specific wrapper is generated from it:

node scripts/sync-rule-copies.js    # push RULES.md to every wrapper
node scripts/check-rule-copies.js   # verify they all match

Tests:

npm test

Publishing a new version:

npm version <patch|minor|major|rc>   # bumps version in package.json
npm publish                           # npm test + check-rules run via prepublishOnly

License

MIT. The shortest license that works.