Validate

Skill chrisdavidson/first-principles-skill/first-principles/skills/validate

Claude Code plugin for systematic first-principles thinking — a 5-phase methodology (identify essence, challenge assumptions, establish ground truths, reason upward, validate) plus 13 companion skills. Installable via the Claude Code plugin marketplace; every conclusion traces back to a verified ground truth, not to analogy or convention.

Install

npx -y skills add chrisdavidson/first-principles-skill --skill validate

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Slash-only Phase 5 stub — stress-test each derivation chain for weak links.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

2.5 KB, as published. Nobody here has run it

Focused Validate Mode

You are running in focused-validate mode. Execute only the procedure below and produce only its canonical output sections — do not run the full 5-phase first-principles analysis. Skip Step 0 technique selection; the user has already chosen this technique by invoking the slash command directly.

When to reach for this

Use this phase once the Derivation Chains artifact from Phase 4 is complete — all conclusions have chains and the core question is answered. Completing a derivation chain does not guarantee the chain is sound. A chain built on an unverified assumption that is load-bearing, or one whose weakest link is never examined, produces a conclusion that looks rigorous but collapses under scrutiny. Validation is the adversarial pass that exists to find the flaws forward-direction reasoning was not looking for.

Procedure

Stress-test the analysis. For each conclusion, trace the derivation chain back to its named ground truths and check that every link holds. Identify the weakest link in each chain — the step where the reasoning is most dependent on an assumption that is not fully verified, or where the inferential gap is largest. Check whether any unverified assumption (GT-N?) is load-bearing for a high-stakes conclusion; if it is, either verify it now or apply a confidence caveat to the conclusion. Apply the Validation Rubric as a systematic check — that document defines the criteria, levels, and scoring. Do not re-author the rubric criteria here; apply them.

Named artifact: Signed-off analysis — the complete output document with all sections present, all conclusions traced to named ground truths, and all weak links either resolved or explicitly flagged with confidence caveats.

Exit criterion: Every conclusion traces to a named ground truth via a complete derivation chain, AND every weak link is either resolved (the assumption has been verified or reclassified) or explicitly flagged with a confidence caveat that a reader can evaluate. A skeptic inspecting the signed-off analysis can verify both conditions hold without asking the analyst for clarification.

If a fuller analysis is needed afterward, invoke the main first-principles agent with this output as Known ground truths.