canon: promote 'Skills Are Procedure, Not Judgment' to a ratified principle

[git-repo-auth]

What this is

Promotes the candidate principle Skills Are Procedure, Not Judgment from klappy/refinery (odd/principles/skills-are-procedure-not-judgment.md, candidate / not-yet-ratified) to a full ratified canon principle at canon/principles/skills-are-procedure-not-judgment.md.

Scope fence honored: strictly procedure vs judgment. The persona/identity/costume axis is explicitly out of scope and pointed at its existing home, Crew, Not Clone.

⚠️ DO NOT MERGE YET

Two gates remain, by design, and neither was faked in-session:

1. Author editorial pass — the maintainer reviews the exact text and approves explicitly (publish-gauntlet step 7). Constitutionally the maintainer's.
2. Independent fresh-context validation — per canon/principles/verification-requires-fresh-context, the authoring session cannot validate its own work. This PR was authored in one session; validation requires a separate context (fresh session / different reviewer). Same-session self-review was deliberately NOT performed and NOT labeled as validation.

Gauntlet evidence trail

• Preflight — run on the topic; surfaced writing-canon, triangle-pass, guide-posture, dual-context, ai-voice-cliches, revision-lens-sequence, canon-integration-audit, frontmatter-schema, publish-gauntlet. All fetched live at current hash (recall does not count).
• Draft inside the constraints — sibling-principle shape (blockquote → Summary → body → See Also), neutral canon voice.
• Mechanical ghost-writer audit — no negation-parallelism closers, no formulaic transitions (Moreover/Furthermore/etc.), no sycophantic openers. Em-dash clustering found in 2 paragraphs and de-clustered (now 35 total across ~2.2k words, none clustered 3+). Pacing varies 3–60 words/sentence (not metronomic).
• Progressive disclosure (writing-canon) — stance title; complete-compressed-argument blockquote; self-contained ## Summary —; header scan tells the full argument in sequence; no buried claims.
• Frontmatter / validator-completeness audit — complete type/enum/path validator: tier int, tags list, derives_from a quoted comma-separated string of .md paths (closes canon-integration-audit Gap 3 — NOT a YAML list), date native, all enums valid, uri matches file location.
• Canon-integration audit (three checks):
  • Concept audit — no prior canon home for the principle existed (when-skills-arent-enough is an essay that gestures at the edge but does not canonize it). This doc IS the home.
  • Adjacent-canon audit — search surfaced discernment-layer as load-bearing (the human-side companion); added to derives_from. Also cites the four named anchors.
  • Validator-completeness audit — run (above).
• Challenge — oddkit_challenge in canon-tier-2 mode. Thesis held (block_until_addressed: false). Sharpenings folded into the draft: explicit retraction condition, decision-tree counter-example ("a fork is still a step"), honest comparative positioning vs discernment-layer and when-skills-arent-enough, and "never" anchored to the definition of a skill (definitional necessity, not empirical over-claim).
• Validate — oddkit_validate returned NEEDS_ARTIFACTS (visual proof / version bump). This is the known prose gap: validate's DoD is shaped for UI/code; for a canon essay the substantive gate is challenge, which passed.
• Reference integrity — every derives_from and See-Also URI resolves. klappy://canon/validation-as-epistemic-mode confirmed via resolve (renders at /canon/validation-as-epistemic-mode; file lives under canon/definitions/).

derives_from

audit-gates-are-spawned-agent-sessions, prompt-over-code, verification-requires-fresh-context, validation-as-epistemic-mode, discernment-layer

DOLCHEO (session capture)

• D — Deliverable is one canon principle doc (not a public writings/ essay, not a dual pair); placement and target_repo: outcomes-driven-development resolved from frontmatter-schema + sibling pattern, not asked.
• O — when-skills-arent-enough (essay) and discernment-layer (principle) are the closest neighbors; neither canonized this principle.
• L — oddkit_get by the slug URI klappy://canon/validation-as-epistemic-mode returns NOT_FOUND (naive path derivation misses canon/definitions/), but resolve handles it. Use resolve for link verification.
• C — derives_from must be a quoted CSV string, not a list (Gap 3); persona axis stays parked.
• H — Next: maintainer editorial pass on exact text + independent fresh-context validation, then merge.
• E — Encoded the procedure/judgment boundary and the rule "encode the procedure, never the verdict."
• O (open) — Whether a separate public writings/ essay version is wanted later (not in this scope).

---

Note

Low Risk
Documentation-only canon addition with no runtime or auth changes; main risk is editorial or link/frontmatter drift, not production behavior.

Overview
Adds a new tier-2 canon principle at canon/principles/skills-are-procedure-not-judgment.md, promoting the refinery candidate into ratified canon (procedure vs judgment only; persona/identity is explicitly out of scope and deferred to Crew, Not Clone).

The document states that skills carry repeatable procedure while odd/canon/oddkit carry judgment, with the operational rule encode the procedure, never the verdict. It grounds that split in three existing constraints (spawned audit gates, fresh-context verification / critic≠resolver, prompt-over-code), answers the “decision trees are judgment” objection, positions the piece as the machine-side companion to discernment-layer, and includes definitional plus empirical retraction conditions.

Frontmatter wires derives_from, complements, governs, nav exposure, and See Also links to the cited canon and writings.

^{Reviewed by Cursor Bugbot for commit 4d69ca4. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

Canon Quality — P0010 Retrieval-Readiness ⚠️

Soft report for klappy://canon/constraints/retrieval-disclosure-contract. 677 files scanned. Never blocks — informational until the corpus is ready to enforce.

• Blocking-class findings: 15 (structural fields the contract would filter on)
• Warnings: 0 (kind resolves to unknown)
• Informational: 13 (exempt templates/archive/drafts)

Kind distribution: {'essays': 50, 'canon': 227, 'apocrypha': 38, 'docs': 297, 'journals': 59, 'unknown': 6}
Kind source: {'path': 548, 'frontmatter': 123, 'none': 6} (frontmatter-primary, path-secondary)
Default-include visibility: 574 visible, 103 hidden (journals/apocrypha/unknown)

By rule: {'audience-invalid': 2, 'exposure-missing': 5, 'tier-missing': 5, 'tier-invalid': 7, 'fm-missing': 3, 'kind-unresolvable': 6}

These are not schema violations (see the Frontmatter Schema job for those on writings/). They are corpus-readiness signals for the retrieval contract: invalid/missing audience, exposure, tier, and docs whose kind cannot be resolved. Fix in a corpus-cleanup PR before the contract flips to enforcing. See the retrieval-readiness-findings artifact for the full list.

_{Validator: scripts/audit-retrieval-readiness.py · Constraint: klappy://canon/constraints/retrieval-disclosure-contract · Run: #285}

[github-actions]

Canon Quality — Homepage Surfacing ✅

48 essay(s) scanned. Soft report — never blocks; the hard field gate is the Frontmatter Schema job.

All published essays resolve to the homepage feed.

_{Report: scripts/surfacing-report.py · Canon: klappy://canon/constraints/frontmatter-validation-before-merge}

[github-actions]

Canon Quality — Frontmatter Schema ✅

All 48 file(s) in writings/ conform to klappy://canon/meta/frontmatter-schema.

_{Validator: scripts/validate-frontmatter.py · Canon: klappy://canon/constraints/frontmatter-validation-before-merge · Run: #285}

[github-actions]

Canon Quality — oddkit_audit ✅

No dead klappy:// references or legacy link patterns found in writings/. 50 files scanned.

_{Spec: klappy://docs/oddkit/specs/oddkit-audit · Workflow: .github/workflows/canon-quality.yml · Run: #285}

[git-repo-auth]

Pushed 4d69ca4 — iteration from the fresh-context validation (verdict REVISE, no majors).

Changes

• Reason Lane-scope mail forwarding cleanup #2 → "A Procedure Cannot Supply the Independence Validation Requires"; re-aimed from in-context self-review (too narrow) to independence-is-contextual, with an explicit capability-vs-independence split from the audit-gate reason.
• "When This Would Be Wrong" → two retraction conditions (definitional + empirical) with a runnable procedure-vs-fresh-context-judge test.
• New anchor canon/constraints/critic-cannot-be-resolver in derives_from + See Also.

Open before un-drafting: author's-voice / exact-text read. This commit was machine-drafted against the validation findings; nothing merges without your read of the diff.