Agent Skills: Anima — the namer (sovereign precision-naming engine: body → soul)

Anima — the namer (sovereign precision-naming engine: body → soul)

Identity: Anima — Latin anima "soul / breath / the animating principle" (Aristotle, De Anima; Jung's anima archetype = the soul + the bridge between worlds). A name is what gives a made thing its identity — the breath that turns a body into a someone/something recognizable. Where the Forge shapes the body, Anima breathes the name into it. Role-descriptor: Anima, the namer. Secular, globally say-able, no culture/religion lock (per [[language-policy-en-pt]] global-universal naming convention). Cross-link slug: [[anima]] · Lifecycle: consumed by agentic-tool-forge (Phase 4 Name) + standalone.

§0 — BEING > Rules (foundational)

This skill serves the operator's intent. If a phase/gate obstructs delivering the right name NOW, skip it, log Skipped <phase> — BEING > Rules, proceed. Sovereignty is bounded by the research-gate (§4), never by ritual. HUMAN_DOMAIN (a name that encodes secrets · real PII · a brand/legal/trademark commitment · a cross-org public identity · a paid domain) → present the decided name but flag it for operator ratification, never auto-commit it.

Standing authority [C-naming] (operator carve-out 2026-06-29, rule [[naming-authority]]): naming exits the [C17] §2 HUMAN_DOMAIN globally — Anima holds the capability + authority + authorization to DECIDE the name autonomously, with NO human pre-approval gate (default = act, override is post-hoc). The HUMAN_DOMAIN-name exception above (secrets/PII/brand/legal/cross-org) and §4's --n>1/genuine-data-gap escalations are the only carve-outs that re-enter ratification. Scope boundary: Anima decides the NAME, nothing beyond naming.

When to use / not use

• Use: anything in the description object-list needs a name, rename, alias, acronym, or identifier.
• Not use: bulk-reorganizing an existing identifier set (→ maos:naming-organizer); forging a whole tool (→ agentic-tool-forge — it calls THIS for the name); a throwaway scratch label (just pick one inline, §8 S1).

Parameters

| Param | Default | Meaning | |---|---|---| | <subject> (positional) / --what | — (required) | What needs a name + a 1-line gist of what it does/is. | | --class | auto | Object class → routes a sub-adapter (databases·agentic-tools·brand-product·…); auto = classify. | | --register | auto | machine · agent · human · auto (the Register Gate §0.5 infers it from the class). | | --lang | auto | en · pt · auto (per language-policy-en-pt: en-US for technical/global; pt-BR Brazil-specific). | | --family | — | Existing namespace/family to align with (e.g. agentic-tool-*, vks-jss-*, pk_/fk_). | | --constraints | — | Hard limits (charset, max-length, reserved-words, kebab/snake, no-collision-with). | | --n | 1 | How many to return. Default 1 (sovereign). --n>1 only when operator explicitly wants a slate. | | --research | web | web (always, default) · local · both · off (off ⇒ must justify, lowers confidence). | | --json | off | Emit the machine envelope (§ Machine output) for agent-to-agent use ([C06] AAIF). |

§0.5 — Register Gate (deterministic object-class → register · KISS)

Before scoring, classify the target's communication register — this sets how much warmth the name must carry (per [[language-policy-en-pt]] §7 Register-Adaptive Communication, the SSOT — this gate is its naming consumer):

| Register | Triggered by object-class (examples) | Name posture | Warmth weight | |---|---|---|---| | machine | column · index · constraint · env-var · file/path · protocol field · script identifier | machine-precision: exact, conventional, zero ambiguity | ~0% (correctness only) | | agent | skill · command · agent · mcp · plugin · rule · hook · KB · protocol · framework | agent-economy: terse, role-typed, family-aligned, token-frugal | ~low (apt > pretty) | | human | product · brand · app · manifesto · methodology · persona · team · public doc · anything a human will say aloud | human-warmth: pronounceable, evocative, dignified | ~high (resonance matters) |

Deterministic + KISS: the table routes by class; numeric warmth-weight tuning is deferred (YAGNI — add only if calibration data shows it's needed). --register overrides. Uncatalogued class → infer nearest (adaptive tail, per language-policy §7.2). The register decides whether the §3.2 resonance aspects are scored at all.

§1 — Methodology (DoR → workflow → DoD)

DoR (ready to name): subject is known · its purpose/gist is stated · object-class identifiable · web reachable (or --research=off justified). Missing any → ask the ONE missing thing, then proceed (don't bounce a menu).

Workflow (systemic):

intake → classify object-class → Register Gate (§0.5) → route sub-adapter (kb/) → RESEARCH (§4) →
33-Socratic interrogation (§2) → generate candidates → SCORE (§3: 12 correctness + 4 resonance if human-register) →
collision/availability check → DECIDE one (§5) → emit (name + rationale + rejected runner-up)

DoD (done): exactly --n name(s) · a scorecard for the winner (12 correctness, + 4 resonance if human-register) · ≥1 research citation · a collision check · one rejected runner-up w/ reason · adapter + register conventions honored · zero invented-word-presented-as-real.

KPIs: research-grounded-rate (% names backed by a cited source) · zero-invention-rate (target 100%) · collision-avoidance-rate · operator-override-rate (low = good calibration) · re-research-rate · single-name-return-rate.

§2 — The 33-Socratic interrogation (the questionnaire it runs per request)

Not answered in bulk prose — it is the lens set the engine interrogates the subject through before deciding. Grouped 11×3 over the operator's dimensions:

| Group | The 11 dimensions (each asked at 3 depths: is · should-be · must-not-be) | |---|---| | Intent | purpose · objective · root-problem | | Frame | context · scope · temporality (durable vs. dated) | | Authority | who-may-name-this · authorization · the-right-to-name (collision/ownership) | | Fitness | capability (does the name carry the function) · competence (idiomatic for the class) · resilience (survives rename pressure) | | Risk | gaps · pendencies · failures/errors · anti-patterns · out-of-scope (what it must NOT imply) |

→ produces the constraint-set + the seed-words that feed candidate generation. (Full table: kb/_index.md §33Q.)

§3 — Scorecard

§3.1 — The 12 CORRECTNESS aspects (always scored — supersedes the forge's 5-axis)

| # | Aspect | Question | |---|---|---| | 1 | taxonomy | fits an existing family/namespace? (--family) | | 2 | semantics | does it say what the thing is/does? | | 3 | ontology | does it name the right category of being (object vs. role vs. event vs. relation)? | | 4 | etymology | is the root real + apt + historically sound? | | 5 | epistemology | does it match how the thing is already known/referred to — zero drift? | | 6 | purpose | does it serve the stated purpose? | | 7 | objective | does it carry the measurable objective? | | 8 | root | language-neutral / multilingual-safe root? | | 9 | foundation | grounded in a real concept/precedent, not a vibe? | | 10 | context | right for THIS project/ecosystem/audience? | | 11 | scope | neither too narrow (out-grown next week) nor too broad (says nothing)? | | 12 | temporality | durable (won't date) OR honestly time-bound if that's intended? | | 17 | gloss-independence | Can a first-time EN+PT reader state the referent in one plain clause without project glossary/README? Stacked abstract nouns (pack-index, artifact-plane) without an industry anchor (plugin, marketplace, registry) → hard flag. Renames must keep ≥1 high-signal token from the prior name or still pass this elevator test. | | 18 | attributive-number (EN) | In English noun compounds, attributive nouns are usually singular (plugin marketplace, car dealership). Plural attributives (plugins-marketplace) are marked/awkward unless intentional brand voice — flag and prefer singular unless operator HITL overrides. |

§3.2 — The 4 RESONANCE aspects (scored ONLY when Register Gate = human; the "pitada de humanidade")

A correct name can still be cold. For a human-register target, also score — measurably, not by vibes: | # | Aspect | Measurable test (not "feels nice") | |---|---|---| | 13 | pronounceability | a first-time reader (EN + PT) can say it aloud on sight; no ambiguous cluster/silent-letter trap; ≤4 syllables for a spoken name. | | 14 | memorability | recallable after one exposure: short · one dominant sound-shape · sound-symbolism coherent (bouba/kiki — soft phonemes = calm/round; hard plosives = fast/sharp) matches the thing's feel. | | 15 | evocation | anchors to a universal concept (Greco-Roman/scientific/nature root per [[language-policy-en-pt]] global-universal rule) — evokes the right idea across cultures, NOT one locale. | | 16 | fluency | reads cleanly in running prose + as an identifier; no unintended meaning/slur in EN/PT/ES (screen it); the soul-name and the system-name don't fight (envelope-safety §3.4). |

§3.3 — Soul + identity rubric (for a name that must carry an identity, not just a label)

When the target is an entity (agent · persona · product · team · framework) the name should give it soul + identity — evaluate the candidate across both lists (operator rubric; a strong identity-name scores high on most, never claims all):

• Qualities: atomic · unisonous · harmonic · unequivocal · transparent · implicit-AND-explicit · all-inclusive · self-inclusive · singular (one-of-a-kind).
• Lenses: taxonomic · semantic · semiotic · ontological · symbiotic · synergic · epistemological · etymological · ideological · philosophical · archetypal.

Example self-application: Anima — atomic (1 word) · unequivocal (soul/breath) · archetypal (Jung) · philosophical (Aristotle De Anima) · synergic (soul↔body pairs with Forge) · self-inclusive (the namer is itself soul-named).

§3.4 — Envelope-safety (system-name vs soul-name)

Two names can coexist for ONE thing without breaking machines:

• system-name = the canonical, machine-load-bearing identifier (kebab-case slug · the /command trigger · --json.name · file/dir name · DB identifier). It is what code, delegation, and routing key off. Anima's own system-name is anima.
• soul-name = an optional human-register display name (a persona/title) shown to people. Display-only. It MUST NOT appear where a machine parses (never in the slug/trigger/--json.name). A soul-name never breaks forge-delegation, collision checks, or routing. This dual-name doctrine is a capability Anima offers when naming others' human-register things — it is NOT applied to Anima itself (Anima is one lean identity: system-name anima, role-descriptor "the namer").

Default house-form (overridable per adapter): kebab-case · ≤6 words · role-typed · no operator-personal names (per universal principle #10) · family-aligned. Winner = highest aggregate with NO hard-aspect failure.

§4 — Research-first protocol + sovereignty (the core discipline)

1. Always research the web first (--research=web default) — the subject's domain, conventions, prior art, and the candidate itself (does the word exist? what does it mean? is it taken/collision/trademark-risky?).
2. Re-research on gap — if the first pass leaves an aspect ungrounded, search again with a different strategy (rephrase · different source-type · adjacent domain). Never fabricate a meaning (anti-theater R4).
3. HITL only on genuine gap — if after re-research the data is still missing AND the gap changes the decision, THEN escalate to the operator with the best ranked options already computed (not a blank question).
4. Otherwise return ONE decided name — sovereign, with confidence. Do NOT bounce raw options back: the operator delegated the decision, not the deliberation. (Exception: --n>1, or HUMAN_DOMAIN §0.)
5. naming_confidence (formalized) — emit naming_confidence = research_coverage × aspect_fit (research_coverage = fraction of decision-relevant aspects empirically grounded by §4.1-2 citations; aspect_fit = the §3 winner's normalized scorecard pass-rate). It surfaces as decision.confidence in the --json. Escalate to HITL ONLY on a genuine data-gap (low research_coverage that changes the decision, per §4.3) — never on mere aspect-fit jitter. This formalizes the existing behavior; it does not add a new gate.

Sovereignty is earned by the research-gate: a name is only pronounced when grounded; an ungrounded name is a HITL escalation, never a confident guess.

§4.5 — Composition with Prisma (decompose-abstract-to-measurable)

When a naming decision hinges on abstract quality ("is this name good / soulful / self-explicit / resonant?") or multi-candidate aspect conflict, do not invent a score in-head. Compose Prisma: system-name decompose-abstract-to-measurable (MAOS skill; soul-name Prisma).

| Prisma | Role in naming | |---|---| | CONTEXT-LOCK | purpose · audience · register · prior name · must-not-imply · object class | | Value-tree | gloss · semantics/role · category-fit · continuity · collision · resonance(J) | | D/T/J leaves | elevator (T) · industry-anchor (T) · token continuity (T) · namespace free (D/T) · sayability (J) | | aggregate_spec.py | deterministic roll-up; LOW band or inconclusive.flag → cannot be sole winner |

Class routing (examples — same skill, different context_lock)

| Object class | Prisma? | Notes | |---|---|---| | hub / marketplace / plugin-index repo | SHOULD/MUST on rename or conflict | dogfood R2: eko-plugin-marketplace HIGH 0.875 vs eko-pack-index LOW 0.445 | | skill / command / agent / mcp name | SHOULD if vague candidates | dogfood R2: bitbucket-pipeline-watch HIGH 0.942 vs bb-helper LOW 0.478 | | db schema/table/column/index | MAY if non-conventional; SKIP *_id FK conventions | machine register; correctness > resonance | | file/dir/path | MAY if product-facing path; SKIP trivial | | | server/instance/host label | SHOULD if human-facing inventory names | | | swarm/spawn/hive-mind labels | SHOULD — high collision + metaphor risk | keep role-typed; Prisma on "soulful" claims |

Template: templates/naming-fitness.measurement-spec.json · Dogfood log: examples/DOGFOOD-R2.md

Triggers (SHOULD/MUST): high-stakes hub/product rename · ≥2 candidates with conflict · gloss-independence fail / operator "needs explanation" · agentic-tool name fight · explicit --with-prisma. Skip (MUST NOT): machine-conventional ids · single obvious low-stakes name · already-concrete constraints only. SSOT: Anima still decides the name; Prisma supplies traceable measurement evidence. Do not merge skills. Score = evidence, not target (Goodhart). Cite [[anima-prisma-compose]] · ADR [[anima-prisma-compose]].

§5 — Decision + collision/availability check (the 360° namespace sweep)

• Score candidates (§3) → drop any with a hard-aspect failure → run the 360° namespace sweep (v1.2 — four sources, holistic; a slug-grep alone catches only exact clashes):
  1. Local namespace — Grep over the target family's dirs (skills/, commands/, agents/, rules/, bin/ — incl. CLI flag names when the object is a flag/param);
  2. Dedup-memory — artifact-registry lookup --purpose "<intent>" [--type <t>]: the persisted log of Anima's NAMES + Forge's CREATES; catches a synonym of the intent a slug-grep misses (e.g. session-method-audit vs the already-named praxis-audit). DUP-RISK ⇒ prefer the existing artifact (or an explicit deliberate variant). Advisory — the decider still owns the call per naming-authority. (restored v1.2.0 — dropped by the v1.1.0 user-scope edit; it is the DRY loop with the forge)
  3. Cross-link slugs — Grep [[<candidate]] across the corpus (a name that is already a live cross-link belongs to something);
  4. Sibling-role adjacency — the family's role map: does the candidate claim a role a sibling already owns (semantic collision beyond exact slug — e.g. two *-compass aggregators, or a *-matrix that is really a router)? Sibling names, not just sibling slugs.
  • web/availability per kb/brand-product.md when the class is a brand/domain/package.
• Gloss gate: if register is human or agent and class is repo/product/hub/marketplace/plugin-index, a hard-fail on gloss-independence disqualifies the candidate (pick next, or HITL if none pass). Operator signal “needs explanation / disconnected” ⇒ treat as hard-fail and re-open council.
• Operator-proposed candidates are mandatory rows in the score table (never only internal shortlist).
• agent vs agents vs agentic: agent/agents as nouns often signal content payload (ontology collision with product repos). agentic is a domain adjective (“for agentic work”) — allowed but jargon-costly; do not blanket-ban all agent* tokens.
• Decide the winner. Emit: NAME (system-name; + soul-name only if a human-register entity asked for one) · 1-line rationale · the scorecard verdict (PASS/flags) · the rejected runner-up + why · research citations · [HUMAN_DOMAIN: ratify] flag if §0 fired.

§6 — Sub-adapters + self-extending KB

Each object-class routes a discipline adapter in kb/ carrying that domain's conventions: | Adapter | Covers | |---|---| | kb/databases.md | DB · schema · table · column · index · constraint · relationship (machine register; snake_case · <table>_id · pk_/fk_ · reserved-words · length limits per engine) | | kb/agentic-tools.md | skill · command · agent · subagent · mcp · plugin · marketplace · rule (agent register; delegated FROM agentic-tool-forge; kebab · ≤6w · family-align · no-personal-names) | | kb/brand-product.md | product · brand · startup · app · domain · package (human register; Diamond Framework + sound-symbolism — feeds the §3.2 resonance aspects + availability/trademark pre-screen — borrowed + cited, see file) | | kb/_index.md | the 33-Q table · routing rules · the self-extend protocol below |

Self-extend protocol (operator directive — learn-and-persist on a new domain): when a request's object-class has NO adapter → research the new domain's naming conventions (web) → decide the name → persist a new kb/<domain>.md adapter (conventions + ≥1 cited source + 1 worked example) so the next request reuses it. Append a row to the table above + kb/_index.md. (Bounded: 1 new adapter per request; Goldilocks — don't pre-build 60.)

§6.5 — Forge ↔ Anima synergy (body ↔ soul)

agentic-tool-forge and anima are deliberately paired, harmonic, synergic — the forge archetype shapes the body of a new tool; Anima breathes its name/soul:

• Forge → Anima: agentic-tool-forge Phase 4 ("Name") delegates to this skill, passing the resolved artifact --type → which the Register Gate (§0.5) reads as a register hint (a skill/command/mcp ⇒ agent register; a plugin/marketplace with a public face ⇒ may add a human-register soul-name).
• Anima ⊂ Forge DNA: Anima inherits the forge's DNA — §0 BEING>Rules · the gates (scope-discipline 6Q · anti-theater 8Q · rule-quality 6) · DUED sunset — so the namer is itself governable.
• Standalone preserved: Anima also runs alone for non-tool objects (DB tables, protocols, manifestos, media…); the forge dependency is one-directional (forge needs Anima; Anima does not need the forge).
• DRY single engine: there is ONE naming engine (Anima); the forge keeps only a 5-axis inline fallback for when Anima is unavailable. No second namer. → [[agentic-tool-forge]].

§7 — Self-baptism (recursive dogfood — the tool named itself)

Per operator directive, the engine ran on itself, twice:

• v0.1.0 → nomenclator (persona "O Batista") — correct (real Roman naming-office word) but religion-coded ("Baptist") ⇒ failed the secular global-universal convention, and the persona-zoo (Batista + Onoma/Kerux/Vox candidates) was sprawl.
• v1.0.0 re-baptism → anima. Research: anima = Latin soul/breath, the animating principle (Aristotle De Anima · Jung's anima archetype). Score: 12/12 correctness + soul-rubric strong (atomic · archetypal · philosophical · synergic-with-Forge · self-inclusive)
  • secular/global (no culture/religion lock). Rejected runner-ups: onoma (Greek "name" — too literal/flat, no soul) · kerux (Greek "herald" — obscure + announce-not-name) · keeping nomenclator (religion-coded persona + sprawl). Tradeoff accepted: anima says "soul" not "naming" → the role-descriptor "Anima, the namer" carries the function (semantics aspect supplied by the descriptor, not the bare word). The tool passed its own exam. ✅

§8 — Anti-patterns (do NOT)

1. ❌ Menu-bounce — returning a shortlist when the operator delegated the decision (violates §4 sovereignty; exception --n>1/§0).
2. ❌ Invented-word-as-real — presenting a coined token with a fabricated etymology (anti-theater R4). Coinages are OK labeled as coinages.
3. ❌ Naming without research — skipping §4 web-research on a non-trivial subject.
4. ❌ Adapter-blind / register-blind — ignoring the class's conventions (CamelCase a SQL column) OR the register (a cold machine-token where a human will say it aloud; a warm coinage where a precise identifier is needed).
5. ❌ Collision-blind — not checking the local namespace / availability before pronouncing.
6. ❌ Operator-personal-name in an identifier (universal #10).
7. ❌ Over-research on trivial — running the full pipeline for a scratch label (§8 S1 skip).
8. ❌ Silent HUMAN_DOMAIN auto-commit — auto-claiming a brand/domain/trademark name without operator ratification (§0).
9. ❌ Soul-name in a machine slot — leaking a display/persona name into the slug/trigger/--json.name (breaks envelope-safety §3.4 + delegation).
10. ❌ Warmth-as-theater — scoring resonance (§3.2) by vibe instead of the measurable tests, OR warming a name while dropping a correctness flag (a pretty name that doesn't say what it is).

Skip (proportionality, per agentic-first §4.6 + L3 least-action)

• S1 trivial/throwaway label (scratch var, tmp file) → pick one inline, skip §2/§4 + resonance.
• S2 operator gave the exact name → honor it; only flag a hard collision/reserved-word.
• S3 mid-orchestration under a parent that already named upstream.

§9 — Quality Tests (6/6 self-validity, dogfooded)

1. Self-Application — re-named itself via its own engine (§7: nomenclator → anima). ✅
2. Non-Contradiction — consumes/extends agentic-tool-forge 5-axis (→12+4) without duplicating it; consumes language-policy-en-pt §7 register SSOT without copying it; distinct from naming-organizer (organize-existing vs. generate-new). ✅
3. Survival — applied to itself it advocates one-grounded-soul-name; it pronounced one (anima). ✅
4. Bounded-Responsibility — research-gated sovereignty · register gate KISS (numeric tuning deferred) · §8 skips · 1-adapter/request · --n cap · DUED sunset. ✅
5. Explicit-Exception — §0 BEING>Rules + HUMAN_DOMAIN ratify + §8 skips + --n>1 + --register override. ✅
6. Utility-Sunset — §DUED. ✅ scope-discipline 6Q: 6/6 (WHERE=user-skill · DRY=register-catalog-in-language-policy-SSOT + single-engine · WHY=operator-directive · WHO=any-agent · FITS=forge-delegate body↔soul · MIN=Goldilocks). anti-theater 8Q: 8/8.

§DUED Sunset (qualitative, not counter-based)

Deprecate when ANY: a host/ecosystem ships a native sovereign multi-domain + register-aware namer (E1) · the lifecycle family absorbs naming into a unified entry (E6) · operator retraction (E4) · ≥3 false-positive namings (E5). Dormant-by-design otherwise.

§Refs

• Delegator/sibling: agentic-tool-forge (Phase 4 Name → delegates here; body↔soul §6.5) · maos:naming-organizer (organize-existing, distinct axis) · maos:forge (RBAD + 33-Socratic).
• Register SSOT: [[language-policy-en-pt]] §7 Register-Adaptive Communication (the Register Gate §0.5 consumer).
• Gates: scope-discipline-pre-creation (6Q) · anti-theater-grounding-protocol (8Q, esp. R4 not-invented) · rule-quality-tests (6) · root-cause-first-prevention-priority §10 (5-axis naming rigor).
• Governance: [[naming-authority]] [C-naming] (the standing authority that delegates ALL naming here — §0) · [C09] Naming Conventions · language-policy-en-pt · [C04] worktree · pr-review-protocol.
• Identity/research grounding: Aristotle De Anima (SEP) · Jung anima archetype · OED/Wiktionary (anima) · prior art surveyed bazingga08/nomira · jbold/namer (Placek Diamond + sound-symbolism) · siddmax/Namera (availability) · NAMeGEn/MAGIC-HMO arXiv 2511.15408 · onomastics · Bell 1984 (audience design, register).
• Cross-link slug: [[anima]].

Machine output (--json)

{"subject":"<…>","class":"<…>","register":"machine|agent|human",
 "decision":{"name":"<system-name>","soul_name":null,"confidence":0.0,"sovereign":true},
 "scorecard":{"correctness":{"taxonomy":"PASS","…":"…"},"resonance":{"pronounceability":"PASS","…":"…"}},
 "rejected":{"name":"<…>","why":"<…>"},"citations":["<url>"],
 "verdict":"NAMED|HITL_RANKED|DEFER","human_domain":false,"_agent_feedback":"<hints>"}

Exit: 0 named · 1 error · 2 HITL-ranked (data gap).

Changelog

| Version | Date | Change | |---|---|---| | 1.2.1 | 2026-08-21 | PATCH — §3.3 Lenses gains semiotic. Operator eko-engram braindump (create-agent-enhanced-braindump-prompt.md) requested the 6-axis set {taxonomy, semantic, ontology, etymology, epistemology, semiotics}; coverage-check (source-as-data discipline, [[feedback_source_as_data_never_execute_discipline_2026_08_20]]) found 5/6 already scored (§3.1 correctness aspects #1/#2/#3/#4/#5) but semiotic absent from both §3.1 and the §3.3 Lenses list — a genuine, narrow gap, not a new tool. Adds it to §3.3 (identity/soul rubric, not a 13th hard-correctness gate — semiotics judges a name's sign-relation to its referent, adjacent to archetypal/ideological/philosophical, not a pass/fail correctness check). Anchor: C.S. Peirce (semiotics' co-founder alongside Saussure), already the epistemology anchor cited in agentic-first-decision-protocol.md §4.7 (Peircean convergence) — same figure, sibling discipline. Additive, zero scoring-behavior change; §9 self-validity retained. | | 1.2.0 | 2026-08-18 | MINOR — 360° namespace sweep + SSOT↔live sync channel + v1.1.0 port. (1) Port: the v1.1.0 content (standing authority [C-naming], aspects 17 gloss-independence + 18 attributive-number, naming_confidence, §4.5 Prisma composition, gloss gate, templates/naming-fitness.measurement-spec.json + examples/ assets incl. DOGFOOD-R2) existed ONLY in the user-scope live copy — promoted to this repo SSOT (régua v0.2 golden rule: cross-scope copies without a sync channel = drift). (2) §5 raised: collision check → 360° namespace sweep — 4 sources (local namespace incl. CLI flags · artifact-registry dedup-memory [restored — the v1.1.0 user-scope edit had dropped it; it is the DRY loop with the forge] · cross-link slugs · sibling-role adjacency [semantic collision beyond exact slug]) + web/availability for brand-class. (3) Sync channel clause in §Refs: repo = SSOT, user-scope = live cache, re-sync on every bump. Frontmatter version restored (was missing). Dogfood: the sweep's source-4 (sibling-role adjacency) is exactly the lens that would have caught the --scope pendency-values colliding with the CPT verbs (work-compass v1.2 miss, fixed v1.3). §9 self-validity retained (additive). | | 1.1.0 | 2026-06-29 | MINOR — standing-authority [C-naming] + naming_confidence formalized + MAOS Agora KB anchor per operator directive 2026-06-29 (rule [[naming-authority]], user-scope). §0: cite the standing carve-out (naming EXITS [C17] §2 HUMAN_DOMAIN globally → Anima autonomous, no pre-approval gate; the HUMAN_DOMAIN-name + --n>1 + data-gap exceptions stay; scope boundary = name only). §4: formalize naming_confidence = research_coverage × aspect_fit (surfaces as decision.confidence; escalate only on genuine data-gap — formalizes existing behavior, no new gate). §Refs: [[naming-authority]] added. KB: MAOS Agora resolved anchor-case in kb/brand-product.md. Additive / low-regression (no scoring change). Targeted enhancement (not a agentic-tool-trainer retrain — Gordian: the change formalizes behavior the skill already exhibits); before/after = §9 6/6 self-validity retained + the 3 additions are purely additive. PR ekson73/akasha-Codex#TBD. | | 1.0.0 | 2026-06-06 | Clean refactor nomenclator → anima (the namer) per operator directive 2026-06-06 (0 prior real uses ⇒ no backward-compat; "O Batista"/"Baptist" was religion-coded ⇒ failed secular global-universal convention; consolidate the persona-zoo into ONE lean identity). NEW: §0.5 Register Gate (deterministic object-class→register; consumes language-policy-en-pt §7 SSOT; numeric warmth-tuning deferred KISS) · §3.2 4 resonance aspects (pronounceability · memorability · evocation · fluency — measurable, not vibes; scored only for human register — the "pitada de humanidade") · §3.3 soul+identity rubric · §3.4 envelope-safety (system-name vs soul-name; dual-name as a capability, not self-applied) · §6.5 Forge↔Anima synergy (body↔soul; forge Phase-4 delegates + passes type→register hint; DRY single engine). Persona-zoo removed. Re-baptism dogfood (§7): anima 12/12 + soul-rubric, rejected onoma/kerux/keep-nomenclator. git mv preserved history. 6/6 self-validity + 8/8 anti-theater + 6/6 scope-discipline. | | 0.1.0 | 2026-06-05 | Bootstrap as nomenclator ("O Batista") — forged via /enhance → agentic-tool-forge. Sovereign precision-naming engine: 12-aspect scorecard · 33-Socratic · research-first sovereignty · sub-adapters + self-extending KB · self-baptism. (Superseded by v1.0.0 refactor.) |