Skip to content

docs(landing): mention v0.3.0 current-gen calibration and link router example #4

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
PBNZ merged 1 commit into from
Apr 20, 2026
Merged

docs(landing): mention v0.3.0 current-gen calibration and link router example #4

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 7 additions & 3 deletions docs/index.html
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -218,11 +218,15 @@ <h2>What makes Newton different from just prompting &ldquo;be more critical&rdqu
<p>&ldquo;Be more critical&rdquo; produces performed scepticism — manufactured objections that sound rigorous but don't change the answer. Newton is calibrated the other way: push back when there's a real reason, agree plainly when you're right, and flag genuine uncertainty rather than either hedging or arguing to seem sharp. The skill spells out what &ldquo;real reason&rdquo; looks like, what self-critique-before-delivery means in practice, and when clarifying questions are load-bearing versus ceremonial — so the behaviour is consistent turn-to-turn rather than drifting with each prompt.</p>
<p>Newton also operationalises reuse-before-rebuild, current-sources research, surgical editing, and handoff-when-scope-outgrows-the-chat as explicit sub-routines the agent runs, not as vibes. That's what makes it a <em>skill</em> rather than a prompt: the agent loads the methodology once and applies it consistently, with visible reuse-check reports and dated sources, rather than re-deriving the disposition from your tone every turn.</p>

<h2>Calibrated for current-generation models</h2>
<p>Newton is model-agnostic by design — the same methodology runs on Opus, Sonnet, and Haiku — but v0.3.0 is tuned so current-generation capabilities (adaptive thinking, implicit tool detection, session memory, parallel execution, high-resolution vision) are leveraged rather than fought. Tool guidance is outcome-oriented instead of prescriptive, independent sub-tasks parallelise, multi-turn sessions can use file-system memory, and visual input is treated as primary evidence. Older models still get the full Newton experience with graceful degradation — no feature relies on a specific model.</p>
Copy link

Copilot AI Apr 20, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The landing page says multi-turn sessions can use file-system memory, but this capability is environment-dependent. To avoid implying it works everywhere, consider adding the same qualifier used in the skill/CHANGELOG (e.g., “when available”).

Suggested change
<p>Newton is model-agnostic by design — the same methodology runs on Opus, Sonnet, and Haiku — but v0.3.0 is tuned so current-generation capabilities (adaptive thinking, implicit tool detection, session memory, parallel execution, high-resolution vision) are leveraged rather than fought. Tool guidance is outcome-oriented instead of prescriptive, independent sub-tasks parallelise, multi-turn sessions can use file-system memory, and visual input is treated as primary evidence. Older models still get the full Newton experience with graceful degradation — no feature relies on a specific model.</p>
<p>Newton is model-agnostic by design — the same methodology runs on Opus, Sonnet, and Haiku — but v0.3.0 is tuned so current-generation capabilities (adaptive thinking, implicit tool detection, session memory, parallel execution, high-resolution vision) are leveraged rather than fought. Tool guidance is outcome-oriented instead of prescriptive, independent sub-tasks parallelise, multi-turn sessions can use file-system memory when available, and visual input is treated as primary evidence. Older models still get the full Newton experience with graceful degradation — no feature relies on a specific model.</p>

Copilot uses AI. Check for mistakes.
<p>Want model-specific behaviour layered on top of <em>every</em> skill, not just Newton? v0.3.0 ships an optional <a href="https://github.com/PBNZ/newton-skill/blob/main/docs/examples/router-claude-md.md">router CLAUDE.md example</a> you can drop into your own <code>~/.claude/CLAUDE.md</code>. It's opt-in — the plugin does not install it for you.</p>
Copy link

Copilot AI Apr 20, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This paragraph is phrased as a v0.3.0-specific claim, but the link targets blob/main, which may drift from the v0.3.0 content over time. To keep the landing page consistent with the version referenced, link to the v0.3.0 tag/relnotes (or otherwise a version-pinned URL).

Suggested change
<p>Want model-specific behaviour layered on top of <em>every</em> skill, not just Newton? v0.3.0 ships an optional <a href="https://github.com/PBNZ/newton-skill/blob/main/docs/examples/router-claude-md.md">router CLAUDE.md example</a> you can drop into your own <code>~/.claude/CLAUDE.md</code>. It's opt-in — the plugin does not install it for you.</p>
<p>Want model-specific behaviour layered on top of <em>every</em> skill, not just Newton? v0.3.0 ships an optional <a href="https://github.com/PBNZ/newton-skill/blob/v0.3.0/docs/examples/router-claude-md.md">router CLAUDE.md example</a> you can drop into your own <code>~/.claude/CLAUDE.md</code>. It's opt-in — the plugin does not install it for you.</p>

Copilot uses AI. Check for mistakes.

<h2>Future development</h2>
<p>Planned after v0.2.0:</p>
<p>Planned after v0.3.0:</p>
<ul>
<li><strong>Additional-language READMEs.</strong> This README is English-only at v0.2.0. Translations (and a landing-page language switcher) are tracked in the <a href="https://github.com/PBNZ/newton-skill/issues">issues</a>.</li>
<li><strong>Sub-agent decomposition (v0.3.0).</strong> Newton's research, reuse-check, and self-critique passes are candidates for dedicated sub-agents — keeping the top-level conversation lean while heavy work happens in focused contexts.</li>
<li><strong>Additional-language READMEs.</strong> The README and this landing page are English-only at v0.3.0. Translations (and a language switcher) are tracked in the <a href="https://github.com/PBNZ/newton-skill/issues">issues</a>.</li>
<li><strong>Sub-agent decomposition.</strong> Newton's research, reuse-check, and self-critique passes are candidates for dedicated sub-agents — keeping the top-level conversation lean while heavy work happens in focused contexts. v0.3.0 shipped a principle-level parallelisation note as the stepping stone; dedicated sub-agents are on the roadmap for a later release.</li>
</ul>
</main>
<footer>
Expand Down
Loading