chore: bump VibeTags to 1.0.0-RC10 and opt into Gemini granular rules

.gemini/rules/se-deversity-skill3-llm-AnthropicChatModel.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for AnthropicChatModel
+
+## Security-Critical Code
+- **Rule**: This code is security-critical. Do not weaken security properties. Every change must be explicitly reviewed for security impact.
+- **Aspect**: Anthropic API credential handling and hosted-provider network egress
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-llm-ChatModel.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for ChatModel
+
+## Contract-Frozen Signature
+- **Constraint**: You may change internal logic, but MUST NOT modify the method name, parameters, return type, or checked exceptions.
+- **Reason**: The single seam every model-driven stage binds to — QueryPlanner, Synthesizer, Verifier and the self-correction Reviser all take this one interface, which is what lets one --llm-provider choice apply uniformly. Test fakes implement it directly, so changing the signature breaks every unit test that avoids a live model.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-llm-LlmProviderFactory.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for LlmProviderFactory
+
+## Security-Critical Code
+- **Rule**: This code is security-critical. Do not weaken security properties. Every change must be explicitly reviewed for security impact.
+- **Aspect**: LLM provider credential resolution and model selection
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-llm-LocalLlmClient.md

@@ -0,0 +1,11 @@
+<!-- VIBETAGS-START -->
+# Rules for LocalLlmClient
+
+### Rules for field apiKey
+- **Rule**: Never log or expose runtime values of this element.
+- **Reason**: LLM provider API key — never log, echo, or include in errors/fixtures
+
+## Security-Critical Code
+- **Rule**: This code is security-critical. Do not weaken security properties. Every change must be explicitly reviewed for security impact.
+- **Aspect**: outbound LLM-provider credential (Bearer token) handling
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-llm-NameSanitizer.md

@@ -0,0 +1,10 @@
+<!-- VIBETAGS-START -->
+# Rules for NameSanitizer
+
+## Security-Critical Code
+- **Rule**: This code is security-critical. Do not weaken security properties. Every change must be explicitly reviewed for security impact.
+- **Aspect**: output sanitization: reserved-word stripping must never be weakened
+
+### Rules for method sanitize
+- **Rule**: Must remain a pure function. Forbid state modifications and side effects.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-llm-SkillMdPostProcessor.md

@@ -0,0 +1,11 @@
+<!-- VIBETAGS-START -->
+# Rules for SkillMdPostProcessor
+
+## Core Functionality
+- **Sensitivity**: High
+- **Note**: Deterministically guarantees SKILL.md spec compliance; model output is never trusted. Changes risk emitting invalid frontmatter — keep the parsing and frontmatter synthesis covered by SkillMdPostProcessorTest.
+
+### Rules for method render
+- **Rule**: This operation is idempotent. Calling it multiple times must produce the same result as calling it once.
+- **Reason**: SelfCorrectionLoop re-runs render() on its own output, so a revised draft passes through repeatedly. Every guarantee here must converge: exactly one frontmatter block and exactly one provenance footer, no matter how many revision rounds ran.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-llm-Verifier.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for Verifier
+
+## Core Functionality
+- **Sensitivity**: High
+- **Note**: Accuracy gate that re-grounds claims against the sources. Only worthwhile with a capable model — a weak model rewrites rather than grounds. Keep the prompt strict about supported-claims-only and announced-vs-shipped.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-model-ContextBundle.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for ContextBundle
+
+## Immutable Type
+- **Rule**: This type is immutable. Never introduce non-final fields, setters, or mutating methods.
+- **Note**: Immutable record; the sources list is defensively copied in the compact constructor.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-model-RunManifest.md

@@ -0,0 +1,6 @@
+<!-- VIBETAGS-START -->
+# Rules for RunManifest
+
+## Schema & Serialization Safety
+- **Rule**: Prohibit altering data formats, fields, database columns, or serialization structures without explicit backward-compatible migration paths.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-model-Source.md

@@ -0,0 +1,10 @@
+<!-- VIBETAGS-START -->
+# Rules for Source
+
+## Architectural Boundary Constraints
+- **Layer**: model
+- **Prohibited References**: se.deversity.skill3.pipeline, se.deversity.skill3.llm, se.deversity.skill3.cli, se.deversity.skill3.skillspector, se.deversity.skill3.web, se.deversity.skill3.net
+
+## Domain Model Boundary
+- **Purity**: Framework-free DDD Entity.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-BraveSearchClient.md

@@ -0,0 +1,11 @@
+<!-- VIBETAGS-START -->
+# Rules for BraveSearchClient
+
+### Rules for field apiKey
+- **Rule**: Never log or expose runtime values of this element.
+- **Reason**: Brave Search subscription token — never log, echo, or include in errors/fixtures
+
+## Security-Critical Code
+- **Rule**: This code is security-critical. Do not weaken security properties. Every change must be explicitly reviewed for security impact.
+- **Aspect**: external-API credential handling and the only network egress with a secret token
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-CutoffResolver.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for CutoffResolver
+
+## Context & Focus
+- **Focus**: Keep the cutoff TABLE small and sourced from published model documentation
+- **Avoid**: hardcoding per-skill logic; the cutoff is always overridable via --cutoff-time
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-DiscoveryProvider.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for DiscoveryProvider
+
+## Security-Critical Code
+- **Rule**: This code is security-critical. Do not weaken security properties. Every change must be explicitly reviewed for security impact.
+- **Aspect**: forwards the Brave subscription token to the search client; must not log it
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-FileCorpus.md

@@ -0,0 +1,8 @@
+<!-- VIBETAGS-START -->
+# Rules for FileCorpus
+
+## Load-Bearing Oddity
+- **Rule**: This looks removable but is deliberate. Refactor only while the invariant holds.
+- **Invariant**: FileCorpus implements BOTH discovery seams — SearchClient and PageFetcher — and LearnCommand injects the same instance into both slots. That is the design, not a layering slip: it is what makes an offline --input-file run take the identical downstream path as a live Brave run, so the two modes cannot diverge.
+- **Breaks if changed**: the class is split into two collaborators, or either interface is dropped — offline runs then follow a different path from live ones and stop proving anything about the real pipeline
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-HttpPageFetcher.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for HttpPageFetcher
+
+## Security-Critical Code
+- **Rule**: This code is security-critical. Do not weaken security properties. Every change must be explicitly reviewed for security impact.
+- **Aspect**: outbound page fetch egress for partly-untrusted URLs; SSRF guard must not be weakened
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-PageFetcher.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for PageFetcher
+
+## Contract-Frozen Signature
+- **Constraint**: You may change internal logic, but MUST NOT modify the method name, parameters, return type, or checked exceptions.
+- **Reason**: Fetch seam. Keeping page retrieval behind it is what lets extraction, date parsing and scoring be tested against HTML fixtures with no network, and it is the boundary at which --input-file replaces the network entirely.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-QueryPlanner.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for QueryPlanner
+
+## Context & Focus
+- **Focus**: keep discovery topic-agnostic — the model plans the queries for any topic
+- **Avoid**: hardcoding per-topic search terms or a fixed query suffix like " documentation"
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-RetrievalService.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for RetrievalService
+
+## Thread-Safety Guarantee
+- **Strategy**: IMMUTABLE
+- **Note**: Collaborators (PageFetcher/HttpClient, DateExtractor, AuthorityScorer) are stateless/immutable; each fetch task builds its own Source and results are merged on the caller thread. Keep it that way — do not share mutable state between fetch tasks. The opt-in `sequential` mode only removes concurrency (fetches run on the caller thread); it cannot weaken the invariant — serial execution is strictly safer than the parallel default it replaces.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-pipeline-SearchClient.md

@@ -0,0 +1,7 @@
+<!-- VIBETAGS-START -->
+# Rules for SearchClient
+
+## Contract-Frozen Signature
+- **Constraint**: You may change internal logic, but MUST NOT modify the method name, parameters, return type, or checked exceptions.
+- **Reason**: Discovery seam. BraveSearchClient (live) and FileCorpus (--input-file) both implement it, and isCuratedCorpus() is what tells the pipeline to skip LLM query planning. Removing the default method, or changing what it returns, silently re-enables planning for a corpus that is already the curated result set.
+<!-- VIBETAGS-END -->

.gemini/rules/se-deversity-skill3-skillspector-InputVetter.md

@@ -0,0 +1,8 @@
+<!-- VIBETAGS-START -->
+# Rules for InputVetter
+
+## Load-Bearing Oddity
+- **Rule**: This looks removable but is deliberate. Refactor only while the invariant holds.
+- **Invariant**: A quarantined source is dropped from the set handed to the synthesizer, but its finding is still recorded and still trips the run gate. Redaction runs FIRST and unconditionally, so a secret never reaches the model even when SkillSpector is unavailable — and when it is unavailable nothing is gated, because absence of findings is observed, never asserted.
+- **Breaks if changed**: quarantining is treated as resolving the finding, redaction is made conditional on the scanner being present, or a skipped scan is reported as clean
+<!-- VIBETAGS-END -->

GEMINI.md

@@ -0,0 +1,62 @@
+# GEMINI.md
+
+AI guardrails for Google Gemini, generated from source annotations by
+[VibeTags](https://github.com/PIsberg/vibetags). The region between the
+VIBETAGS-START and VIBETAGS-END markers is regenerated on every compile;
+never hand-edit inside it. Per-element detail lives in ,
+indexed from the block below.
+
+<!-- VIBETAGS-START -->
+# AUTO-GENERATED AI RULES
+# Generated by VibeTags | https://github.com/PIsberg/vibetags
+# Do not edit manually.
+
+## 🔒 PII / PRIVACY GUARDRAILS
+The following elements handle Personally Identifiable Information (PII).
+NEVER include their runtime values in logs, console output, external API calls,
+test fixtures, mock data, or code suggestions.
+
+- `se.deversity.skill3.llm.LocalLlmClient.apiKey`: LLM provider API key — never log, echo, or include in errors/fixtures
+- `se.deversity.skill3.pipeline.BraveSearchClient.apiKey`: Brave Search subscription token — never log, echo, or include in errors/fixtures
+
+## 🧠 CORE FUNCTIONALITY (CHANGE WITH EXTREME CAUTION)
+The following elements are well-tested core components. Make changes with extreme caution.
+
+- `se.deversity.skill3.llm.SkillMdPostProcessor`: Sensitivity: High. Note: Deterministically guarantees SKILL.md spec compliance; model output is never trusted. Changes risk emitting invalid frontmatter — keep the parsing and frontmatter synthesis covered by SkillMdPostProcessorTest.
+- `se.deversity.skill3.llm.Verifier`: Sensitivity: High. Note: Accuracy gate that re-grounds claims against the sources. Only worthwhile with a capable model — a weak model rewrites rather than grounds. Keep the prompt strict about supported-claims-only and announced-vs-shipped.
+
+## 🔐 SECURITY-CRITICAL CODE
+The following elements are security-critical. AI must not weaken security properties. Any change must be reviewed for security impact.
+
+- `se.deversity.skill3.llm.AnthropicChatModel`: Security-critical code [Anthropic API credential handling and hosted-provider network egress]. Do not weaken security properties. Flag any change for security review.
+- `se.deversity.skill3.llm.LlmProviderFactory`: Security-critical code [LLM provider credential resolution and model selection]. Do not weaken security properties. Flag any change for security review.
+- `se.deversity.skill3.llm.LocalLlmClient`: Security-critical code [outbound LLM-provider credential (Bearer token) handling]. Do not weaken security properties. Flag any change for security review.
+- `se.deversity.skill3.llm.NameSanitizer`: Security-critical code [output sanitization: reserved-word stripping must never be weakened]. Do not weaken security properties. Flag any change for security review.
+- `se.deversity.skill3.pipeline.BraveSearchClient`: Security-critical code [external-API credential handling and the only network egress with a secret token]. Do not weaken security properties. Flag any change for security review.
+- `se.deversity.skill3.pipeline.DiscoveryProvider`: Security-critical code [forwards the Brave subscription token to the search client; must not log it]. Do not weaken security properties. Flag any change for security review.
+- `se.deversity.skill3.pipeline.HttpPageFetcher`: Security-critical code [outbound page fetch egress for partly-untrusted URLs; SSRF guard must not be weakened]. Do not weaken security properties. Flag any change for security review.
+
+## Scoped Rules Index
+Detailed per-element guardrails live in scoped rule files that load automatically when you open the matching source file. Consult the referenced file before modifying an element:
+
+- `se.deversity.skill3.llm.AnthropicChatModel` → `.gemini/rules/se-deversity-skill3-llm-AnthropicChatModel.md`
+- `se.deversity.skill3.llm.ChatModel` → `.gemini/rules/se-deversity-skill3-llm-ChatModel.md`
+- `se.deversity.skill3.llm.LlmProviderFactory` → `.gemini/rules/se-deversity-skill3-llm-LlmProviderFactory.md`
+- `se.deversity.skill3.llm.LocalLlmClient` → `.gemini/rules/se-deversity-skill3-llm-LocalLlmClient.md`
+- `se.deversity.skill3.llm.NameSanitizer` → `.gemini/rules/se-deversity-skill3-llm-NameSanitizer.md`
+- `se.deversity.skill3.llm.SkillMdPostProcessor` → `.gemini/rules/se-deversity-skill3-llm-SkillMdPostProcessor.md`
+- `se.deversity.skill3.llm.Verifier` → `.gemini/rules/se-deversity-skill3-llm-Verifier.md`
+- `se.deversity.skill3.model.ContextBundle` → `.gemini/rules/se-deversity-skill3-model-ContextBundle.md`
+- `se.deversity.skill3.model.RunManifest` → `.gemini/rules/se-deversity-skill3-model-RunManifest.md`
+- `se.deversity.skill3.model.Source` → `.gemini/rules/se-deversity-skill3-model-Source.md`
+- `se.deversity.skill3.pipeline.BraveSearchClient` → `.gemini/rules/se-deversity-skill3-pipeline-BraveSearchClient.md`
+- `se.deversity.skill3.pipeline.CutoffResolver` → `.gemini/rules/se-deversity-skill3-pipeline-CutoffResolver.md`
+- `se.deversity.skill3.pipeline.DiscoveryProvider` → `.gemini/rules/se-deversity-skill3-pipeline-DiscoveryProvider.md`
+- `se.deversity.skill3.pipeline.FileCorpus` → `.gemini/rules/se-deversity-skill3-pipeline-FileCorpus.md`
+- `se.deversity.skill3.pipeline.HttpPageFetcher` → `.gemini/rules/se-deversity-skill3-pipeline-HttpPageFetcher.md`
+- `se.deversity.skill3.pipeline.PageFetcher` → `.gemini/rules/se-deversity-skill3-pipeline-PageFetcher.md`
+- `se.deversity.skill3.pipeline.QueryPlanner` → `.gemini/rules/se-deversity-skill3-pipeline-QueryPlanner.md`
+- `se.deversity.skill3.pipeline.RetrievalService` → `.gemini/rules/se-deversity-skill3-pipeline-RetrievalService.md`
+- `se.deversity.skill3.pipeline.SearchClient` → `.gemini/rules/se-deversity-skill3-pipeline-SearchClient.md`
+- `se.deversity.skill3.skillspector.InputVetter` → `.gemini/rules/se-deversity-skill3-skillspector-InputVetter.md`
+<!-- VIBETAGS-END -->

build.gradle

@@ -12,7 +12,7 @@ plugins {
 group = 'se.deversity'
 version = '0.1.0'
 
-ext.vibetagsVersion = '1.0.0-RC8'
+ext.vibetagsVersion = '1.0.0-RC10'
 // code-karta renders the diagrams under docs/diagrams straight from this source tree.
 ext.codekartaVersion = '0.2.0'