nccgs 2.1.1 → 3.0.0-rc.2

package/.claude/rules/nccgs-networking.md

@@ -1,13 +0,0 @@
----
-paths:
-  - "Assets/**/*Network*.cs"
-  - "Assets/**/*Multiplayer*.cs"
----
-
-# NCCGS networking rules
-
-- Treat authority, protocol, message version, persistence, and service changes as
-  CONTROLLED unless project policy explicitly narrows the risk.
-- Validate untrusted input and failure paths; make latency, reconnect, and rollback
-  assumptions testable.
-- Preserve compatibility unless an approved migration says otherwise.

package/.claude/rules/nccgs-performance.md

@@ -1,14 +0,0 @@
----
-paths:
-  - "Assets/**/*.cs"
-  - "Assets/**/*.shader"
-  - "Assets/**/*.hlsl"
----
-
-# NCCGS performance rules
-
-- Use project budgets and reproducible target-platform scenarios.
-- Measure before broad optimization; compare identical before/after conditions.
-- Treat demonstrated hot-path allocation, loading, frame pacing, GPU, and memory
-  failures as evidence, not generic style rules.
-- Preserve correctness, determinism, and visual/product targets while optimizing.

package/.claude/rules/nccgs-rendering.md

@@ -1,16 +0,0 @@
----
-paths:
-  - "Assets/**/*.shader"
-  - "Assets/**/*.hlsl"
-  - "Assets/**/*Shader*"
-  - "Assets/**/*VFX*"
----
-
-# NCCGS rendering rules
-
-- Respect the configured render pipeline, target platforms, visual direction, and
-  measured GPU budgets.
-- Record visual changes as reproducible scenarios and leave subjective acceptance to
-  the Product Owner.
-- Render-pipeline, package, and broad project-setting changes are CONTROLLED.
-- Use official Unity rendering skills for API and Editor operation mechanics.

package/.claude/rules/nccgs-security.md

@@ -1,13 +0,0 @@
----
-paths:
-  - "Assets/**/*.cs"
-  - "Packages/**/*.json"
----
-
-# NCCGS security rules
-
-- Never embed secrets, credentials, signing material, or privileged endpoints.
-- Treat client state and network/service input as untrusted at declared boundaries.
-- Dependency, analytics, commerce, account, privacy, and external-service changes
-  require explicit authority and targeted security evidence.
-- Do not expose sensitive values in logs or evidence records.

package/.claude/rules/nccgs-tests.md

@@ -1,18 +0,0 @@
----
-paths:
-  - "Assets/**/Tests/**/*.cs"
-  - "Tests/**/*.cs"
----
-
-# NCCGS test rules
-
-- Map tests to behavior or requirement IDs, not implementation trivia.
-- Bug fixes require a regression test when the failure can be reproduced
-  deterministically.
-- Keep EditMode tests independent of scene/runtime state unless integration is the
-  purpose; use PlayMode or runtime verification for behavior that requires Unity.
-- Cover the normal path and material edge cases. Do not inflate coverage with
-  assertions that do not prove the acceptance criterion.
-- A green test run is objective evidence, not subjective gameplay validation.
-- Follow project test conventions before choosing new folders, assemblies, helpers,
-  or frameworks.

package/.claude/rules/nccgs-ui.md

@@ -1,14 +0,0 @@
----
-paths:
-  - "Assets/**/*UI*.cs"
-  - "Assets/**/*.uxml"
-  - "Assets/**/*.uss"
----
-
-# NCCGS Unity UI rules
-
-- UI presents and requests state; authoritative gameplay state remains outside views.
-- Follow approved UX, input, localization, and accessibility contracts.
-- Verify navigation/focus, errors, scaling, layout expansion, binding lifecycle, and
-  platform input rather than checking appearance only.
-- Use official Unity UI skills for operation and API mechanics.

package/.claude/rules/nccgs-unity-assets.md

@@ -1,21 +0,0 @@
----
-paths:
-  - "Assets/**/*.unity"
-  - "Assets/**/*.prefab"
-  - "Assets/**/*.asset"
-  - "Assets/**/*.meta"
-  - "ProjectSettings/**"
----
-
-# NCCGS Unity asset and settings rules
-
-- Read the ownership section of `.nccgs/project.yaml` before modifying assets or
-  settings.
-- If a connected Editor can perform the operation, use the appropriate official
-  Unity Skill and Unity CLI/Pipeline; do not hand-edit serialized YAML.
-- Inspect before modifying and make the smallest targeted change. Preserve existing
-  hierarchy, GUIDs, serialized references, import settings, and handcrafted values
-  unless the request requires changing them.
-- Package, service, platform, render-pipeline, input-system, and compatibility
-  changes require the authority specified in project policy.
-- Verify persistent Editor changes after saving and reimporting where applicable.

package/.claude/rules/nccgs-unity-code.md

@@ -1,22 +0,0 @@
----
-paths:
-  - "Assets/**/*.cs"
-  - "Packages/**/*.cs"
----
-
-# NCCGS Unity code rules
-
-- Follow `.nccgs/project.yaml`, governing decisions, and established project
-  patterns before generic recommendations.
-- Preserve approved behavior and serialized compatibility unless the task explicitly
-  authorizes a change.
-- Do not introduce packages, services, architecture patterns, or global managers
-  merely because they are common Unity practices.
-- Separate testable domain logic from presentation when it materially improves
-  verification; do not force an abstraction for trivial code.
-- Avoid allocations in demonstrated hot paths, but profile before broad performance
-  rewrites.
-- Treat warnings, nullable assumptions, Unity object lifetime, domain reload, and
-  serialization behavior as part of correctness.
-- Use the relevant official Unity Skill for Unity-specific mechanics rather than
-  relying on copied NCCGS guidance.

package/.claude/skills/accessibility-review/SKILL.md

@@ -1,22 +0,0 @@
----
-name: accessibility-review
-description: Reviews a Unity feature, screen, control scheme, content, or build for scoped accessibility barriers and testable accommodations.
-allowed-tools: Read, Glob, Grep, Agent, WebSearch, WebFetch
----
-
-# Accessibility Review
-
-Resolve execution policy from .nccgs/execution.json before model routing. Missing
-execution.json means compatibility; model profile alone never removes obligations.
-
-- standalone: load the [shared standalone procedure](../../nccgs/protocols/standalone.md)
-  for this workflow, using its intake, mode, bounded domain steps and report rules.
-  Do not load the compatibility branches or inherit their external approval chain.
-- compatibility: read models.profile from .nccgs/project.yaml and load exactly one:
-  astra-claude uses [Astra/Claude procedure](references/astra.md); quality, balanced
-  or inherit uses [Legacy procedure](references/legacy.md). Existing task records
-  retain their original gates even after project execution policy changes.
-
-Unknown execution policies or model profiles are configuration errors. External
-review in standalone is off, advisory or required according to execution policy;
-configured routes do not prove actual provider model or effort.

package/.claude/skills/accessibility-review/references/astra.md

@@ -1,17 +0,0 @@
-# Accessibility Review — astra-claude
-
-Read [the pipeline protocol](../../../nccgs/protocols/astra-claude.md) and
-[generated routing](../../../nccgs/routing.generated.md). Use Astra's supplied plan,
-requirements, and exact context references. Never substitute a legacy procedure.
-
-Domain objective: Reviews a Unity feature, screen, control scheme, content, or build for scoped accessibility barriers and testable accommodations.
-
-Work only inside the supplied plan and assigned write set. Activate an optional
-specialist with the agents command only when its domain is required. Keep one writer
-per path and use official Unity Skills for Editor/assets/runtime operations.
-
-This procedure is read-only: return evidence-backed findings and decisions to Astra
-or the implementer. Supplemental analysis does not replace the mandatory Fable gate.
-
-Use [task gate commands](../../../nccgs/protocols/task-gates.md) for evidence and
-state transitions. Report outcome, evidence, pending decisions, and next owner.

package/.claude/skills/accessibility-review/references/legacy.md

@@ -1,7 +0,0 @@
-# Accessibility Review
-
-Read `.nccgs/project.yaml`, `.nccgs/state.md`, and the active context packet when present. Preserve Product Owner authority and unrelated work.
-
-Remain read-only unless explicitly asked to implement. Read project accessibility targets, UX/design, platforms, inputs, content, and implementation. Review applicable vision, hearing, motor, cognitive, motion, timing, language, and difficulty-support barriers. Report severity, affected task/player, evidence, requirement, feasible remediation, and validation scenario. Do not claim compliance or lived-experience validation.
-
-Return outcome, evidence, affected records, blockers or decisions, residual risk, and the exact next workflow. Do not claim DONE unless this is the `closure` skill and every configured gate passes.

package/.claude/skills/architecture-decision/SKILL.md

@@ -1,22 +0,0 @@
----
-name: architecture-decision
-description: Analyzes and records a focused Unity architecture decision when system boundaries, compatibility, dependencies, persistence, networking, or migration risk require durable rationale.
-allowed-tools: Read, Glob, Grep, Write, Edit, Bash, Agent, AskUserQuestion, WebSearch, WebFetch
----
-
-# Architecture Decision
-
-Resolve execution policy from .nccgs/execution.json before model routing. Missing
-execution.json means compatibility; model profile alone never removes obligations.
-
-- standalone: load the [shared standalone procedure](../../nccgs/protocols/standalone.md)
-  for this workflow, using its intake, mode, bounded domain steps and report rules.
-  Do not load the compatibility branches or inherit their external approval chain.
-- compatibility: read models.profile from .nccgs/project.yaml and load exactly one:
-  astra-claude uses [Astra/Claude procedure](references/astra.md); quality, balanced
-  or inherit uses [Legacy procedure](references/legacy.md). Existing task records
-  retain their original gates even after project execution policy changes.
-
-Unknown execution policies or model profiles are configuration errors. External
-review in standalone is off, advisory or required according to execution policy;
-configured routes do not prove actual provider model or effort.

package/.claude/skills/architecture-decision/references/astra.md

@@ -1,12 +0,0 @@
-# Architecture Decision — astra-claude
-
-Read [the pipeline protocol](../../../nccgs/protocols/astra-claude.md) and
-[generated routing](../../../nccgs/routing.generated.md). Use Astra's supplied plan,
-requirements, and exact context references. Never substitute a legacy procedure.
-
-Astra owns this workflow and its canonical outputs. Claude only inspects relevant
-inputs and returns factual constraints, missing information, or proposed amendments.
-Do not create a replacement plan/document or sign off as Astra.
-
-Use [task gate commands](../../../nccgs/protocols/task-gates.md) for evidence and
-state transitions. Report outcome, evidence, pending decisions, and next owner.

package/.claude/skills/architecture-decision/references/legacy.md

@@ -1,7 +0,0 @@
-# Architecture Decision
-
-Read `.nccgs/project.yaml`, `.nccgs/state.md`, and the active context packet when present. Preserve Product Owner authority and unrelated work.
-
-Read policy, context packet, affected contracts, code, tests, and existing decisions. Define forces and invariants; present materially distinct options with consequences; recommend one; obtain Product Owner authority for dependencies, compatibility, services, or irreversible migration. Record the accepted decision using the project template, including status, scope, alternatives, consequences, implementation constraints, verification, rollback, and superseded decisions. Do not create an ADR for a private local implementation choice.
-
-Return outcome, evidence, affected records, blockers or decisions, residual risk, and the exact next workflow. Do not claim DONE unless this is the `closure` skill and every configured gate passes.

package/.claude/skills/asset-audit/SKILL.md

@@ -1,22 +0,0 @@
----
-name: asset-audit
-description: Audits Unity source assets, import settings, naming, references, size/budget, duplication, ownership, and content-pipeline integrity without mass-editing them.
-allowed-tools: Read, Glob, Grep, Bash, Agent
----
-
-# Asset Audit
-
-Resolve execution policy from .nccgs/execution.json before model routing. Missing
-execution.json means compatibility; model profile alone never removes obligations.
-
-- standalone: load the [shared standalone procedure](../../nccgs/protocols/standalone.md)
-  for this workflow, using its intake, mode, bounded domain steps and report rules.
-  Do not load the compatibility branches or inherit their external approval chain.
-- compatibility: read models.profile from .nccgs/project.yaml and load exactly one:
-  astra-claude uses [Astra/Claude procedure](references/astra.md); quality, balanced
-  or inherit uses [Legacy procedure](references/legacy.md). Existing task records
-  retain their original gates even after project execution policy changes.
-
-Unknown execution policies or model profiles are configuration errors. External
-review in standalone is off, advisory or required according to execution policy;
-configured routes do not prove actual provider model or effort.

package/.claude/skills/asset-audit/references/astra.md

@@ -1,17 +0,0 @@
-# Asset Audit — astra-claude
-
-Read [the pipeline protocol](../../../nccgs/protocols/astra-claude.md) and
-[generated routing](../../../nccgs/routing.generated.md). Use Astra's supplied plan,
-requirements, and exact context references. Never substitute a legacy procedure.
-
-Domain objective: Audits Unity source assets, import settings, naming, references, size/budget, duplication, ownership, and content-pipeline integrity without mass-editing them.
-
-Work only inside the supplied plan and assigned write set. Activate an optional
-specialist with the agents command only when its domain is required. Keep one writer
-per path and use official Unity Skills for Editor/assets/runtime operations.
-
-This procedure is read-only: return evidence-backed findings and decisions to Astra
-or the implementer. Supplemental analysis does not replace the mandatory Fable gate.
-
-Use [task gate commands](../../../nccgs/protocols/task-gates.md) for evidence and
-state transitions. Report outcome, evidence, pending decisions, and next owner.

package/.claude/skills/asset-audit/references/legacy.md

@@ -1,7 +0,0 @@
-# Asset Audit
-
-Read `.nccgs/project.yaml`, `.nccgs/state.md`, and the active context packet when present. Preserve Product Owner authority and unrelated work.
-
-Remain read-only. Read asset policy, protected/user-owned settings, budgets, source/import paths, and generated-path rules. Use official Unity Skills for asset and serialized-reference inspection. Report missing/broken references, import drift, budget violations, duplicates, naming issues that affect tooling, and unsafe ownership. Separate source asset problems from generated/import artifacts; do not normalize hand-tuned settings automatically.
-
-Return outcome, evidence, affected records, blockers or decisions, residual risk, and the exact next workflow. Do not claim DONE unless this is the `closure` skill and every configured gate passes.

package/.claude/skills/audit/SKILL.md

@@ -1,22 +0,0 @@
----
-name: audit
-description: Performs a read-only audit of an NCCGS Unity project, workflow, codebase, design, architecture, ownership, security, performance, evidence, or migration state. Use for diagnosis and gap analysis without implementing changes.
-allowed-tools: Read, Glob, Grep, Bash, Agent, WebSearch, WebFetch
----
-
-# Audit
-
-Resolve execution policy from .nccgs/execution.json before model routing. Missing
-execution.json means compatibility; model profile alone never removes obligations.
-
-- standalone: load the [shared standalone procedure](../../nccgs/protocols/standalone.md)
-  for this workflow, using its intake, mode, bounded domain steps and report rules.
-  Do not load the compatibility branches or inherit their external approval chain.
-- compatibility: read models.profile from .nccgs/project.yaml and load exactly one:
-  astra-claude uses [Astra/Claude procedure](references/astra.md); quality, balanced
-  or inherit uses [Legacy procedure](references/legacy.md). Existing task records
-  retain their original gates even after project execution policy changes.
-
-Unknown execution policies or model profiles are configuration errors. External
-review in standalone is off, advisory or required according to execution policy;
-configured routes do not prove actual provider model or effort.

package/.claude/skills/audit/references/astra.md

@@ -1,17 +0,0 @@
-# Audit — astra-claude
-
-Read [the pipeline protocol](../../../nccgs/protocols/astra-claude.md) and
-[generated routing](../../../nccgs/routing.generated.md). Use Astra's supplied plan,
-requirements, and exact context references. Never substitute a legacy procedure.
-
-Domain objective: Performs a read-only audit of an NCCGS Unity project, workflow, codebase, design, architecture, ownership, security, performance, evidence, or migration state. Use for diagnosis and gap analysis without implementing changes.
-
-Work only inside the supplied plan and assigned write set. Activate an optional
-specialist with the agents command only when its domain is required. Keep one writer
-per path and use official Unity Skills for Editor/assets/runtime operations.
-
-This procedure is read-only: return evidence-backed findings and decisions to Astra
-or the implementer. Supplemental analysis does not replace the mandatory Fable gate.
-
-Use [task gate commands](../../../nccgs/protocols/task-gates.md) for evidence and
-state transitions. Report outcome, evidence, pending decisions, and next owner.

package/.claude/skills/audit/references/dimensions.md

@@ -1,46 +0,0 @@
-# Audit dimensions
-
-Select only dimensions relevant to the request.
-
-## Project and governance
-
-- NCCGS installation integrity and project-policy completeness
-- canonical-source map and authority conflicts
-- protected/user-owned content boundaries
-- compact state accuracy and stale open decisions
-- closure mode, canonical record, linked artifacts, waivers, and legacy safeguards
-
-## Design and traceability
-
-- implementability and internal consistency
-- requirement IDs for meaningful contracts
-- objective acceptance versus subjective gates
-- code/design drift and downstream change impact
-
-## Architecture and code
-
-- module/data-flow boundaries and dependency direction
-- serialization, save, network, and package/service contracts
-- project conventions versus accidental legacy patterns
-- correctness, lifecycle, error handling, performance evidence, and testability
-
-## Verification
-
-- requirement-to-evidence coverage
-- EditMode, PlayMode, runtime, build, and regression adequacy
-- claims based only on file existence or inspection
-- subjective gates incorrectly marked technical PASS
-
-## Unity integration
-
-- Unity version and project configuration
-- official Unity Skills presence and integrity
-- Unity CLI/Pipeline availability and connected Editor behavior
-- hand-edited serialized assets where live tooling should be used
-- package/service changes lacking authority
-
-## Release and operations
-
-- reproducible targets and version identity
-- release-blocking tests, migrations, known issues, and rollback readiness
-- secrets, signing, external publishing, and approval boundaries

package/.claude/skills/audit/references/legacy.md

@@ -1,9 +0,0 @@
-# Audit
-
-Remain read-only. Define scope, authority, expected standards, and evidence before inspecting. Read policy and only the canonical sources and artifacts relevant to the requested audit.
-
-Select applicable dimensions from [audit dimensions](../references/dimensions.md). Route independent lenses to specialists when their conclusions can be compared without overlapping writes. For broad or CONTROLLED audits, use the adversarial reviewer to challenge assumptions and evidence.
-
-Every finding must identify severity, exact evidence, violated contract or concrete failure mode, impact, confidence, and smallest remediation direction. Separate confirmed defects, risks, missing evidence, and optional improvements. Do not convert style preference into a finding.
-
-Report strengths worth preserving, prioritized findings, systemic causes, decision blockers, and an ordered remediation plan. Do not edit files or silently transition into implementation.

package/.claude/skills/balance-review/SKILL.md

@@ -1,22 +0,0 @@
----
-name: balance-review
-description: Analyzes mechanics, economy, combat, progression, rewards, and tuning data for outliers, degenerate strategies, fairness risks, and sensitivity without changing approved balance.
-allowed-tools: Read, Glob, Grep, Bash, Agent
----
-
-# Balance Review
-
-Resolve execution policy from .nccgs/execution.json before model routing. Missing
-execution.json means compatibility; model profile alone never removes obligations.
-
-- standalone: load the [shared standalone procedure](../../nccgs/protocols/standalone.md)
-  for this workflow, using its intake, mode, bounded domain steps and report rules.
-  Do not load the compatibility branches or inherit their external approval chain.
-- compatibility: read models.profile from .nccgs/project.yaml and load exactly one:
-  astra-claude uses [Astra/Claude procedure](references/astra.md); quality, balanced
-  or inherit uses [Legacy procedure](references/legacy.md). Existing task records
-  retain their original gates even after project execution policy changes.
-
-Unknown execution policies or model profiles are configuration errors. External
-review in standalone is off, advisory or required according to execution policy;
-configured routes do not prove actual provider model or effort.

package/.claude/skills/balance-review/references/astra.md

@@ -1,17 +0,0 @@
-# Balance Review — astra-claude
-
-Read [the pipeline protocol](../../../nccgs/protocols/astra-claude.md) and
-[generated routing](../../../nccgs/routing.generated.md). Use Astra's supplied plan,
-requirements, and exact context references. Never substitute a legacy procedure.
-
-Domain objective: Analyzes mechanics, economy, combat, progression, rewards, and tuning data for outliers, degenerate strategies, fairness risks, and sensitivity without changing approved balance.
-
-Work only inside the supplied plan and assigned write set. Activate an optional
-specialist with the agents command only when its domain is required. Keep one writer
-per path and use official Unity Skills for Editor/assets/runtime operations.
-
-This procedure is read-only: return evidence-backed findings and decisions to Astra
-or the implementer. Supplemental analysis does not replace the mandatory Fable gate.
-
-Use [task gate commands](../../../nccgs/protocols/task-gates.md) for evidence and
-state transitions. Report outcome, evidence, pending decisions, and next owner.

package/.claude/skills/balance-review/references/legacy.md

@@ -1,7 +0,0 @@
-# Balance Review
-
-Read `.nccgs/project.yaml`, `.nccgs/state.md`, and the active context packet when present. Preserve Product Owner authority and unrelated work.
-
-Read canonical formulas, tuning values, goals, telemetry/playtest evidence, and interacting systems. Recalculate assumptions, inspect boundaries and long-run behavior, compare player strategies, and identify dominant or dead choices. Separate mathematical defects from subjective target choices. Return evidence, sensitivity, options, recommendation, and PO gates; do not write new balance values unless implementation is explicitly requested.
-
-Return outcome, evidence, affected records, blockers or decisions, residual risk, and the exact next workflow. Do not claim DONE unless this is the `closure` skill and every configured gate passes.

package/.claude/skills/bug-triage/SKILL.md

@@ -1,22 +0,0 @@
----
-name: bug-triage
-description: Reproduces, classifies, prioritizes, and routes Unity defects using player impact, data risk, scope, evidence, ownership, and regression exposure.
-allowed-tools: Read, Glob, Grep, Write, Edit, Bash, Agent
----
-
-# Bug Triage
-
-Resolve execution policy from .nccgs/execution.json before model routing. Missing
-execution.json means compatibility; model profile alone never removes obligations.
-
-- standalone: load the [shared standalone procedure](../../nccgs/protocols/standalone.md)
-  for this workflow, using its intake, mode, bounded domain steps and report rules.
-  Do not load the compatibility branches or inherit their external approval chain.
-- compatibility: read models.profile from .nccgs/project.yaml and load exactly one:
-  astra-claude uses [Astra/Claude procedure](references/astra.md); quality, balanced
-  or inherit uses [Legacy procedure](references/legacy.md). Existing task records
-  retain their original gates even after project execution policy changes.
-
-Unknown execution policies or model profiles are configuration errors. External
-review in standalone is off, advisory or required according to execution policy;
-configured routes do not prove actual provider model or effort.

package/.claude/skills/bug-triage/references/astra.md

@@ -1,27 +0,0 @@
-# Bug Triage — astra-claude
-
-Read [the pipeline protocol](../../../nccgs/protocols/astra-claude.md) and
-[generated routing](../../../nccgs/routing.generated.md). Use Astra's supplied plan,
-requirements, and exact context references. Never substitute a legacy procedure.
-
-Domain objective: Reproduces, classifies, prioritizes, and routes Unity defects using player impact, data risk, scope, evidence, ownership, and regression exposure.
-
-Work only inside the supplied plan and assigned write set. Activate an optional
-specialist with the agents command only when its domain is required. Keep one writer
-per path and use official Unity Skills for Editor/assets/runtime operations.
-
-For code changes: task init with Astra's plan and --risk FAST|STANDARD|CONTROLLED
-(default STANDARD), bind the session, and use the recorded implementation route.
-Keep one accountable implementer; use the scout only for missing facts. Implement,
-run relevant checks, and run task check --id ID --target review before requesting
-independent Fable review. Keep only one Fable review in flight for the task. Fix actionable findings
-within scope and re-review only within the remaining task budget. Re-recording
-implementation does not reset that budget. Exhaustion without a usable PASS means
-stop with evidence for an explicit external decision, never automatic approval.
-After task check --target report passes, write and record the report for Astra.
-Stop at AWAITING_ASTRA_CHECK. Astra checking and explicit Product Owner acceptance
-happen outside the Claude implementation session. FAST uses concise plan/report
-artifacts and keeps the full acceptance chain.
-
-Use [task gate commands](../../../nccgs/protocols/task-gates.md) for evidence and
-state transitions. Report outcome, evidence, pending decisions, and next owner.

package/.claude/skills/bug-triage/references/legacy.md

@@ -1,7 +0,0 @@
-# Bug Triage
-
-Read `.nccgs/project.yaml`, `.nccgs/state.md`, and the active context packet when present. Preserve Product Owner authority and unrelated work.
-
-Read the report, logs, affected requirements, recent changes, and project severity policy. Reproduce through the strongest safe path; distinguish product defect, test defect, environment issue, data corruption, and unknown. Record severity, frequency, scope, evidence, suspected ownership without claiming unproven cause, containment, regression needs, and next role. Escalate data-loss, security, save, network, or release-critical defects to CONTROLLED.
-
-Return outcome, evidence, affected records, blockers or decisions, residual risk, and the exact next workflow. Do not claim DONE unless this is the `closure` skill and every configured gate passes.