dotfiles/claude/memory/pending-spec-corpus-audit-2026-05-14.md

---
name: pending-jurist-directive-arc-specification-corpus-audit
description: Full audit brief issued by jurist 2026-05-14, forwarded by steward 2026-05-15
  evening. DEFERRED to a dedicated half-day session — the rigor is the deliverable,
  not speed. Read-only investigation producing a structured report; no edits, no commits.
  Verbatim copy preserved below.
metadata:
  node_type: memory
  type: project
  originSessionId: 8b47c0eb-d396-43d5-98eb-7623e6f04357
permalink: claude-memory/pending-spec-corpus-audit-2026-05-14
---

## Status

- **Issued by jurist:** 2026-05-14
- **Forwarded by steward:** 2026-05-15 (evening, after §2/§3 multilingual-spec sign-off and v0.5 commit `9cf40d9`)
- **Status:** DEFERRED to a dedicated session
- **Reason for deferral:** The work is genuinely deep (6 specs end-to-end + cross-doc consistency + stale-content detection + gap identification + extraction-candidate assessment + structured report with priority ranking). Realistic effort ~90–180 min of focused analytical reading. The audit's value is the whole-corpus view, which deserves its own attention rather than being fitted in after a substantive production day. Mine completing first will also let the audit verify cross-reference recall quality — an implicit benefit not in the original brief.

## Output target (when executed)

Per the brief: a structured markdown report delivered to the jurist interface for review. Severity markers: `[INCONSISTENCY]`, `[STALE]`, `[EXTRACTION CANDIDATE]`, `[GAP]`, `[INTEGRATION NEEDED]`, `[MINOR]`. Closes with priority-5 ranking. **No edits to specs. No commits.** Audit produces findings only.

Suggested file location for the report when written: `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/_audits/spec-corpus-audit-2026-05-XX.md` (creating `_audits/` if absent).

## Specs in scope

- `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/AldineXXI-specification.md` — framework spine
- `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/content-typology-specification.md` — 14 of 14 §1 types locked; §2 apparatus matrix not yet authored
- `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/vignette-specification.md` — complete, jurist-approved
- `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/essay-versioning-specification.md` — complete mechanism spec
- `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/frontmatter-specification.md` — current; §2.6 language fields recently added
- `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/multilingual-typography-specification.md` — v0.5 at `9cf40d9` (§1 French operational, §2 Spanish deepened, §3 English composed); the brief was issued when this was at v0.1 ("§1 French complete; §2 Spanish pending") so audit task 5 needs updated state-of-spec inputs

---

## Verbatim brief (jurist, 2026-05-14)

# Specification Corpus Audit -- Executor Brief

**Issued by:** Jurist (Claude.app) **Date:** 2026-05-14 **For:** Claude Code **Output:** A written audit report delivered to the jurist interface for review before any changes are made. This audit produces findings only -- no edits, no commits.
* * *

## Context

The ARC specification corpus has grown organically across multiple sessions. The following documents currently exist:

- `AldineXXI-specification.md` -- framework spine
- `content-typology-specification.md` -- 14 of 14 §1 types locked; §2 apparatus matrix not yet authored
- `vignette-specification.md` -- complete, jurist-approved
- `essay-versioning-specification.md` -- complete mechanism spec
- `frontmatter-specification.md` -- current; §2.6 language fields recently added
- `multilingual-typography-specification.md` -- §1 French complete; §2 Spanish pending

This audit assesses the corpus as a whole: internal consistency, extraction candidates, gaps, and stale content.

* * *

## Audit tasks

### 1. Standalone spec candidates

For each section currently inside `AldineXXI-specification.md`, assess whether it has reached standalone-spec weight using these criteria:

- Is it cross-referenced by two or more other documents or systems?
- Does it have its own internal structure (subsections, tables, rules) that would benefit from a dedicated header and status block?
- Would CC or a skill agent need to consult it independently of the AldineXXI framework document?
- Is it actively growing or likely to grow in ways that would make it unwieldy inside AldineXXI?

Report each candidate with: section name, current location, extraction rationale, proposed standalone filename, and what AldineXXI would retain (typically a summary and pointer).

### 2. Stale content within existing specs

Scan each standalone spec document for:

- Sections or statements contradicted by later decisions recorded in changelogs or other specs
- "Proposed authoring sequence" or similar planning sections that are now partially or fully complete and need updating
- Open questions that have since been resolved elsewhere but not propagated back
- Status labels (LOCKED / DRAFT / OPEN) that no longer reflect current state
- Cross-references to other documents that use outdated section numbers, filenames, or status descriptions

For `content-typology-specification.md` specifically: the changelog entries from 2026-04-23 contain several decisions that appear to be recorded but not yet integrated into the body of the spec (Chamber-as-apparatus resolution, deepened ✦ definition, "humic references" preamble). Flag each such decision with its current location (changelog) and its target location (the body section it belongs in).

### 3. Cross-document consistency

Check the following specific consistency questions:

- **`class:` enumeration**: does the set of `class:` values in `frontmatter-specification.md` §2.1 match the final catalogue in `content-typology-specification.md` §1? List any divergences.
- **Vignette applicability**: does the list of types receiving vignettes in `vignette-specification.md` match the applicability rules in `AldineXXI-specification.md` §IX and the content typology catalogue? New types (Heteronym, Hommage, Tombeau, Dream, Reverie, Gloss) were added after the vignette spec was written -- their vignette eligibility has not been explicitly declared.
- **Versioning eligibility**: does `essay-versioning-specification.md` correctly describe which types version? The spec predates the full typology. Check whether the "Zibaldone-form" framing (observations, glimpses) still accurately captures the no-versioning class, given that Fragments, Dreams, Reveries, and other types have since been added.
- **Reading Compass glyphs**: compile the full glyph table from `content-typology-specification.md` and verify it against `AldineXXI-specification.md` §VI. Flag any mismatches, including the Photo Essay glyph reassignment (⊙ → ◧) and the addition of ☾ (Dream) and other new types.
- **Chamber**: now resolved as §2 apparatus, not a §1 type. Verify that all references to Chamber as a content type (in AldineXXI §V, frontmatter spec §2.1, and any other location) are flagged for update.
- **Ornament retirement**: the 2026-04-23 changelog notes a 7→2 ornament retirement. Verify the current ornament count in `AldineXXI-specification.md` §IV against whatever the post-retirement state should be, and flag if the spec body has not been updated.

### 4. Gap identification

Identify specification gaps -- systems or decisions that are referenced or implied across the corpus but have no governing document. For each gap, report: what system is ungoverned, where it is referenced, and what downstream consumers are blocked by the absence.

The following are known gaps from jurist review; confirm whether each remains open and identify any additional gaps not on this list:

- Apparatus specification (extraction from AldineXXI §VII; required before §2 apparatus binding matrix can be authored)
- Silence taxonomy specification (currently embedded in AldineXXI §II; cross-referenced implicitly by apparatus binding)
- Navigation philosophy document (Reading Compass semantics and enfilade logic; referenced in content typology relationship table but not yet written)
- ARC Style Guide (referenced in content typology relationship table; skeleton produced in an earlier session -- locate and assess status)
- §2 apparatus binding matrix itself (the primary outstanding work in content typology)

### 5. The content typology "proposed authoring sequence"

The sequence at the bottom of `content-typology-specification.md` was written when the spec was at an earlier stage. Steps 1 and 2 are now complete. Produce a revised authoring sequence reflecting current state, noting which steps are done, which are in progress, and which remain.

* * *

## Output format

Deliver the audit as a structured markdown report with one section per audit task above. Within each section, use a flat list of findings -- each finding on its own line with a severity marker:
- `[INCONSISTENCY]` -- two documents contradict each other
- `[STALE]` -- content no longer reflects current decisions
- `[EXTRACTION CANDIDATE]` -- section ready or nearly ready for standalone spec
- `[GAP]` -- ungoverned system blocking downstream work
- `[INTEGRATION NEEDED]` -- decision recorded in changelog, not yet in spec body
- `[MINOR]` -- small fix, low consequence

Close the report with a **Priority ranking**: the five findings whose resolution would most unblock downstream work, in order.

Do not make any edits. Do not commit anything. Deliver the report to the jurist interface.

---

## Notes for tomorrow's executor

- The brief was issued when multilingual-typography-spec was at v0.1. Today's work brought it to v0.5 with §2 Spanish deepened and §3 English composed (commit `9cf40d9` on ARC main). Audit task 5's "revised authoring sequence" must reflect this updated state, not the brief's snapshot.
- The chamber-typography mine should be complete by tomorrow morning. The audit can include a recall-quality test against the new palace as an implicit-but-load-bearing finding: do the specs' cross-references actually retrieve coherently when the palace is queried for spec topics?
- Phase 6 mine validation is the natural prerequisite to the audit being grounded — without a queryable palace, some of the audit's verification is harder.
- The audit is read-only. Memory load: substantial but bounded (6 spec files, all in `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/`). Consider using `Agent` with a long-form analyst persona for the cross-document consistency tasks if context conservation is needed.
- The output report when written: `~/_Dev/animal-davidglidden-eu/docs/AldineXXI-Codex/_audits/spec-corpus-audit-2026-05-XX.md` (create `_audits/` if absent).