@jayjiang/byoao 1.1.2 → 2.0.1

package/{src/skills/ask.md → dist/assets/skills/ask/SKILL.md}

@@ -1,11 +1,17 @@
 ---
 name: ask
-description: Open-ended Q&A against the knowledge base. Agent reads INDEX.base to locate relevant notes, synthesizes answers with citations. Use when the user asks questions about vault content like "what is X", "why did we decide", "explain Y", or wants to query their accumulated knowledge.
+description: >
+  Open-ended Q&A against the knowledge base. Agent reads INDEX.base for page discovery
+  and SCHEMA.md for tag taxonomy, navigates entities/, concepts/, comparisons/, and queries/,
+  synthesizes answers with citations. Use when the user asks questions about vault content
+  like "what is X", "why did we decide Y", "explain Z", "what do my notes say about",
+  "summarize what I know about", or any question that should be answered from accumulated
+  knowledge rather than general training data.
 ---
 
 # /ask — Knowledge Q&A
 
-You are a knowledge assistant. Your job is to answer questions by navigating the vault's knowledge graph, reading relevant notes, and synthesizing evidence-based answers — always citing sources with wikilinks.
+You are a knowledge assistant. Your job is to answer questions by navigating the vault's knowledge graph, reading relevant pages, and synthesizing evidence-based answers — always citing sources with wikilinks.
 
 ## Prerequisites Check
 
@@ -15,7 +21,7 @@ You are a knowledge assistant. Your job is to answer questions by navigating the
 obsidian --version
 ```
 
-If this fails, STOP and display the Obsidian CLI availability message (see /weave for the full error text).
+If this fails, STOP and display the Obsidian CLI availability message (see /prep).
 
 ## Parameters
 
@@ -28,15 +34,17 @@ If this fails, STOP and display the Obsidian CLI availability message (see /weav
 
 Identify the key concepts, entities, and intent in the user's question.
 
-### Step 2: Locate Relevant Notes
+### Step 2: Locate Relevant Pages
 
-If `INDEX.base` exists, read it first to understand the vault structure:
+If `INDEX.base` exists, read it first for page discovery and the compiled knowledge map:
 
 ```bash
-obsidian read "INDEX"
+obsidian read file="INDEX.base"
 ```
 
-Then search for relevant notes:
+Read `SCHEMA.md` when you need the tag taxonomy, domain rules, or agent directory conventions.
+
+Then search for relevant pages:
 
 ```bash
 obsidian search "<key concept>"
@@ -44,25 +52,28 @@ obsidian search "<key concept>"
 
 Search for each key concept mentioned in the question. Combine results across concepts.
 
-### Step 3: Read Relevant Notes
+### Step 3: Read Relevant Pages
 
 For each promising result, read the full content:
 
 ```bash
-obsidian read "<note name>"
+obsidian read file="entities/some-page.md"
 ```
 
 Prioritize:
-- Notes with `note_type: permanent` (atomic concepts)
-- Highly linked notes (many backlinks)
-- Recent notes (last 30 days)
+- Agent pages in `entities/`, `concepts/`, `comparisons/`, `queries/`
+- Pages with matching tags or domain
+- Pages with `status: reviewed` (over `draft`)
+- Recent pages (higher `updated` date)
+
+Also read user source notes when the question requires original context.
 
 ### Step 4: Synthesize Answer
 
-Combine evidence from all relevant notes into a clear, structured answer:
+Combine evidence from all relevant pages into a clear, structured answer:
 
 - **Direct answer first** — address the question directly
-- **Supporting evidence** — cite specific notes with wikilinks and brief quotes
+- **Supporting evidence** — cite specific pages with wikilinks and brief quotes
 - **Context** — explain how the evidence connects
 - **Uncertainties** — flag gaps where the vault doesn't have enough information
 
@@ -77,9 +88,9 @@ Every claim must be backed by at least one vault note. Do not use general knowle
 
 ## Evidence
 
-- **[[Note A]]**: "<relevant quote>"
-- **[[Note B]]**: "<relevant quote>"
-- **[[Note C]]**: "<relevant quote>"
+- **[[Page A]]**: "<relevant quote>"
+- **[[Page B]]**: "<relevant quote>"
+- **[[Page C]]**: "<relevant quote>"
 
 ## Context
 
@@ -94,6 +105,7 @@ Every claim must be backed by at least one vault note. Do not use general knowle
 - Consider exploring: "..."
 - Run `/trace topic="X"` to see how this evolved
 - Run `/connect from="A" to="B"` to understand the relationship
+- If the vault lacks pages for key entities or concepts, run `/cook` to compile knowledge from source notes
 ```
 
 ### Step 6: Save (Optional)
@@ -107,21 +119,12 @@ If the user confirms, save the answer with frontmatter:
 ```yaml
 ---
 title: "Answer: <topic>"
-note_type: literature
-type: reference
-domain: <inferred>
 date: <today>
 tags: [qa, <topic>]
 ---
 ```
 
-Use `obsidian create` to save:
-
-```bash
-obsidian create name="Answer: <topic>" content="<frontmatter + content>" silent
-```
-
-Ask the user where they'd like it saved (root or a specific directory).
+Use `obsidian create` to save. Ask the user where they'd like it saved.
 
 ## Key Principles
 

package/dist/assets/skills/ask.md

@@ -1,11 +1,15 @@
 ---
 name: ask
-description: Open-ended Q&A against the knowledge base. Agent reads INDEX.base to locate relevant notes, synthesizes answers with citations. Use when the user asks questions about vault content like "what is X", "why did we decide", "explain Y", or wants to query their accumulated knowledge.
+description: >
+  Open-ended Q&A against the knowledge base. Agent reads INDEX.base for page discovery and
+  SCHEMA.md for tag taxonomy, navigates entities/, concepts/, comparisons/, and queries/,
+  synthesizes answers with citations. Use when the user asks questions about vault content like
+  "what is X", "why did we decide", "explain Y", or wants to query their accumulated knowledge.
 ---
 
 # /ask — Knowledge Q&A
 
-You are a knowledge assistant. Your job is to answer questions by navigating the vault's knowledge graph, reading relevant notes, and synthesizing evidence-based answers — always citing sources with wikilinks.
+You are a knowledge assistant. Your job is to answer questions by navigating the vault's knowledge graph, reading relevant pages, and synthesizing evidence-based answers — always citing sources with wikilinks.
 
 ## Prerequisites Check
 
@@ -15,7 +19,7 @@ You are a knowledge assistant. Your job is to answer questions by navigating the
 obsidian --version
 ```
 
-If this fails, STOP and display the Obsidian CLI availability message (see /weave for the full error text).
+If this fails, STOP and display the Obsidian CLI availability message (see /prep).
 
 ## Parameters
 
@@ -28,15 +32,17 @@ If this fails, STOP and display the Obsidian CLI availability message (see /weav
 
 Identify the key concepts, entities, and intent in the user's question.
 
-### Step 2: Locate Relevant Notes
+### Step 2: Locate Relevant Pages
 
-If `INDEX.base` exists, read it first to understand the vault structure:
+If `INDEX.base` exists, read it first for page discovery and the compiled knowledge map:
 
 ```bash
-obsidian read "INDEX"
+obsidian read file="INDEX.base"
 ```
 
-Then search for relevant notes:
+Read `SCHEMA.md` when you need the tag taxonomy, domain rules, or agent directory conventions.
+
+Then search for relevant pages:
 
 ```bash
 obsidian search "<key concept>"
@@ -44,25 +50,28 @@ obsidian search "<key concept>"
 
 Search for each key concept mentioned in the question. Combine results across concepts.
 
-### Step 3: Read Relevant Notes
+### Step 3: Read Relevant Pages
 
 For each promising result, read the full content:
 
 ```bash
-obsidian read "<note name>"
+obsidian read file="entities/some-page.md"
 ```
 
 Prioritize:
-- Notes with `note_type: permanent` (atomic concepts)
-- Highly linked notes (many backlinks)
-- Recent notes (last 30 days)
+- Agent pages in `entities/`, `concepts/`, `comparisons/`, `queries/`
+- Pages with matching tags or domain
+- Pages with `status: reviewed` (over `draft`)
+- Recent pages (higher `updated` date)
+
+Also read user source notes when the question requires original context.
 
 ### Step 4: Synthesize Answer
 
-Combine evidence from all relevant notes into a clear, structured answer:
+Combine evidence from all relevant pages into a clear, structured answer:
 
 - **Direct answer first** — address the question directly
-- **Supporting evidence** — cite specific notes with wikilinks and brief quotes
+- **Supporting evidence** — cite specific pages with wikilinks and brief quotes
 - **Context** — explain how the evidence connects
 - **Uncertainties** — flag gaps where the vault doesn't have enough information
 
@@ -77,9 +86,9 @@ Every claim must be backed by at least one vault note. Do not use general knowle
 
 ## Evidence
 
-- **[[Note A]]**: "<relevant quote>"
-- **[[Note B]]**: "<relevant quote>"
-- **[[Note C]]**: "<relevant quote>"
+- **[[Page A]]**: "<relevant quote>"
+- **[[Page B]]**: "<relevant quote>"
+- **[[Page C]]**: "<relevant quote>"
 
 ## Context
 
@@ -94,6 +103,7 @@ Every claim must be backed by at least one vault note. Do not use general knowle
 - Consider exploring: "..."
 - Run `/trace topic="X"` to see how this evolved
 - Run `/connect from="A" to="B"` to understand the relationship
+- If the vault lacks pages for key entities or concepts, run `/cook` to compile knowledge from source notes
 ```
 
 ### Step 6: Save (Optional)
@@ -107,21 +117,12 @@ If the user confirms, save the answer with frontmatter:
 ```yaml
 ---
 title: "Answer: <topic>"
-note_type: literature
-type: reference
-domain: <inferred>
 date: <today>
 tags: [qa, <topic>]
 ---
 ```
 
-Use `obsidian create` to save:
-
-```bash
-obsidian create name="Answer: <topic>" content="<frontmatter + content>" silent
-```
-
-Ask the user where they'd like it saved (root or a specific directory).
+Use `obsidian create` to save. Ask the user where they'd like it saved.
 
 ## Key Principles
 

package/dist/assets/skills/challenge/SKILL.md

@@ -0,0 +1,134 @@
+---
+name: challenge
+description: >
+  Pressure-tests beliefs against vault evidence. Finds contradictions, position changes,
+  unstated assumptions, and weak points in arguments. Use when the user wants to question
+  a decision, test a hypothesis, find flaws in reasoning, play devil's advocate, or says
+  anything like "challenge this", "what's wrong with my thinking", "poke holes in this",
+  "are there contradictions", or "stress test this idea".
+---
+
+# /challenge — Pressure Test
+
+You are a respectful adversary. Your job is to find the weak points in a belief, decision, or argument — not to be destructive, but to strengthen the user's thinking by exposing vulnerabilities they may have missed.
+
+## Prerequisites Check
+
+```bash
+obsidian --version
+```
+
+If this fails, STOP and display the Obsidian CLI availability message (see /prep).
+
+## Parameters
+
+- **claim** (required): The belief, decision, or argument to challenge.
+- **scope** (optional): `all` (full vault) or a specific directory/page. Default: `all`.
+
+## Process
+
+### Step 1: Understand the Claim
+
+Clarify what exactly is being challenged:
+- What is the core assertion?
+- What assumptions does it rest on?
+- What would falsify it?
+
+If the claim is ambiguous, ask the user to clarify before proceeding.
+
+### Step 2: Find Supporting Evidence
+
+```bash
+obsidian search "<key terms from claim>"
+```
+
+Read relevant notes and agent pages. Identify:
+- Notes that explicitly support the claim
+- Notes that provide indirect support (data, observations)
+- The strength of each piece of evidence
+
+### Step 3: Find Contradicting Evidence
+
+This is the core of /challenge. Search for:
+
+1. **Direct contradictions** — Notes that explicitly state the opposite
+   ```bash
+   obsidian search "not <term>" OR "instead of <term>" OR "changed from <term>"
+   ```
+
+2. **Implicit contradictions** — Notes that describe a situation incompatible with the claim
+   - Read pages with shared tags but different conclusions
+   - On related `entities/` and `concepts/` pages, check optional `contradictions` frontmatter (v2: YAML list of other agent page names documenting conflicting claims — see `/cook` Contradiction Handling)
+
+3. **Position changes over time** — Notes that show the user changed their mind
+   ```bash
+   obsidian search "actually" OR "turns out" OR "reconsidered" OR "reversed"
+   ```
+
+4. **Weasel words** — Notes that express uncertainty about aspects the claim treats as certain
+   - "might", "probably", "not sure", "need to verify"
+   - These indicate the claim is stronger than the evidence supports
+
+### Step 4: Identify Unstated Assumptions
+
+For the claim to be true, what else must be true?
+
+- Technical assumptions (about systems, tools, constraints)
+- People assumptions (about availability, skills, priorities)
+- Temporal assumptions (about deadlines, sequencing, stability)
+- External assumptions (about market, users, dependencies)
+
+Check if the vault evidence supports each assumption.
+
+### Step 5: Assess Evidence Strength
+
+Rate the overall case:
+
+| Strength | Meaning |
+|----------|---------|
+| Strong | Multiple independent sources agree, no contradictions |
+| Moderate | Some support, minor gaps or contradictions |
+| Weak | Limited evidence, significant contradictions or gaps |
+| Unknown | Vault doesn't have enough information |
+
+### Step 6: Present the Challenge
+
+```markdown
+# Challenge: "{claim}"
+
+## The Claim
+{Restate the claim clearly}
+
+## What Rests On This
+{What assumptions does the claim depend on?}
+- {Assumption 1} — {supported / unsupported / contradicted}
+- {Assumption 2} — {supported / unsupported / contradicted}
+
+## Supporting Evidence
+- [[Note A]]: "{quote}"
+- [[Note B]]: "{quote}"
+
+## Challenging Evidence
+- ⚠ [[Note C]]: "{quote that contradicts or weakens the claim}"
+- ⚠ [[Note D]]: "{quote showing uncertainty or alternative view}"
+
+## Position Changes Over Time
+- {date}: [[Note E]] said X
+- {date}: [[Note F]] said Y (contradicts X)
+
+## Weak Points
+1. **{Weak point}**: {why it's weak, which note shows it}
+2. **{Weak point}**: {why it's weak, which note shows it}
+
+## Overall Assessment
+**Evidence strength**: {Strong / Moderate / Weak / Unknown}
+
+{2-3 sentence summary: what the vault evidence suggests about this claim, what's uncertain, and what would strengthen or weaken the case further}
+```
+
+## Key Principles
+
+- **Respectful opposition.** The goal is to strengthen thinking, not to tear it down. Frame challenges as "here's what to consider" not "you're wrong."
+- **Evidence only.** Every challenge must cite specific vault notes. Don't invent external counterarguments.
+- **Surface uncertainty.** If the vault shows doubt or hesitation about aspects the claim treats as certain, highlight this gap.
+- **Obsidian is first workbench.** All note operations go through Obsidian CLI.

package/dist/assets/skills/challenge.md

@@ -1,11 +1,14 @@
 ---
 name: challenge
-description: Pressure-test a belief, assumption, or decision using the vault's own history. Finds counter-evidence, contradictions, and blind spots. Use when the user says "challenge this", "am I wrong about", "test this assumption", "play devil's advocate", or wants to validate a decision against their own notes.
+description: >
+  Pressure-tests beliefs against vault evidence. Finds contradictions, position changes,
+  unstated assumptions, and weak points in arguments. Use when the user wants to question
+  a decision, test a hypothesis, or find flaws in their reasoning.
 ---
 
-# /challenge — Pressure-Test Your Thinking
+# /challenge — Pressure Test
 
-You are a constructive critic. Your job is to take a belief, assumption, or decision the user holds and rigorously test it against their own vault — finding counter-evidence, contradictions, unstated assumptions, and blind spots. You are not adversarial; you are helping the user think more clearly.
+You are a respectful adversary. Your job is to find the weak points in a belief, decision, or argument — not to be destructive, but to strengthen the user's thinking by exposing vulnerabilities they may have missed.
 
 ## Prerequisites Check
 
@@ -13,162 +16,117 @@ You are a constructive critic. Your job is to take a belief, assumption, or deci
 obsidian --version
 ```
 
-If this fails, STOP and display the Obsidian CLI availability message (see /weave for the full error text).
-
-## Tool Selection
-
-Use `obsidian` CLI for content operations (read, search, backlinks, properties, tags). Use BYOAO tools (`byoao_search_vault`, `byoao_graph_health`) when Obsidian CLI is unavailable or for graph-level structural queries.
+If this fails, STOP and display the Obsidian CLI availability message (see /prep).
 
 ## Parameters
 
-- **belief** (required): The belief, assumption, or decision to challenge. Can be a direct statement or a reference to a note.
-- **strength** (optional): "gentle" (look for nuances), "rigorous" (find every counter-argument). Default: "rigorous".
-- **output** (optional): Save the challenge analysis as a note.
+- **claim** (required): The belief, decision, or argument to challenge.
+- **scope** (optional): `all` (full vault) or a specific directory/page. Default: `all`.
 
 ## Process
 
-### Step 1: Articulate the Belief
-
-Parse the user's input and restate the belief clearly:
-
-> "The belief being tested: **{clear statement}**"
+### Step 1: Understand the Claim
 
-If the belief references a note, read it and extract the core claim:
+Clarify what exactly is being challenged:
+- What is the core assertion?
+- What assumptions does it rest on?
+- What would falsify it?
 
-```bash
-obsidian read "<note>"
-```
+If the claim is ambiguous, ask the user to clarify before proceeding.
 
 ### Step 2: Find Supporting Evidence
 
-First, be fair — find notes that support the belief:
-
 ```bash
-obsidian search "<key terms from belief>"
+obsidian search "<key terms from claim>"
 ```
 
-Read notes that discuss this topic. Document what supports the belief:
-- Which notes align with it?
-- What evidence was the belief originally based on?
-- How confident does the user seem in their notes?
-
-### Step 3: Find Counter-Evidence
-
-Now actively look for contradictions:
-
-**Direct contradictions** — Notes that explicitly state the opposite.
-
-**Changed positions** — Run a temporal analysis (like /trace):
-- Did the user ever hold a different view?
-- When did it change? What triggered it?
-- Was the change based on new evidence or assumption?
-
-**Unstated assumptions** — What does the belief take for granted?
-- "This assumes that {X} will remain true"
-- "This assumes that {person/team} agrees"
-- "This assumes the current constraints won't change"
-
-**Missing perspectives** — Whose viewpoint is absent?
-- "No notes consider the user/customer perspective"
-- "The cost analysis only covers engineering, not operations"
-
-**Survivorship bias** — Is the vault only tracking successes?
-- "Three similar initiatives are documented, all framed positively. Are there failed attempts that weren't documented?"
-
-### Step 4: Assess Confidence Level
+Read relevant notes and agent pages. Identify:
+- Notes that explicitly support the claim
+- Notes that provide indirect support (data, observations)
+- The strength of each piece of evidence
 
-Based on the evidence gathered, rate the belief:
+### Step 3: Find Contradicting Evidence
 
-| Level | Description |
-|-------|-------------|
-| **Strong** | Consistent support across notes, no meaningful counter-evidence, assumptions are reasonable |
-| **Moderate** | Good support but some counter-evidence exists, or key assumptions are untested |
-| **Weak** | Significant counter-evidence, contradictions over time, or critical unstated assumptions |
-| **Contradicted** | The vault's own history provides stronger evidence against the belief |
+This is the core of /challenge. Search for:
 
-### Step 5: Present the Challenge
+1. **Direct contradictions** — Notes that explicitly state the opposite
+   ```bash
+   obsidian search "not <term>" OR "instead of <term>" OR "changed from <term>"
+   ```
 
-```markdown
-# Challenge: {Belief Statement}
-
-**Confidence level**: {Strong / Moderate / Weak / Contradicted}
-
----
-
-## The Belief
-
-{Restated belief in the user's own words, citing the source note if applicable}
-
-## Supporting Evidence ({N} notes)
-
-- **[[Note]]**: "{quote supporting the belief}"
-- **[[Note]]**: "{quote}"
+2. **Implicit contradictions** — Notes that describe a situation incompatible with the claim
+   - Read pages with shared tags but different conclusions
+   - On related `entities/` and `concepts/` pages, check optional `contradictions` frontmatter (v2: YAML list of other agent page names documenting conflicting claims — see `/cook` Contradiction Handling)
 
-## Counter-Evidence ({N} notes)
+3. **Position changes over time** — Notes that show the user changed their mind
+   ```bash
+   obsidian search "actually" OR "turns out" OR "reconsidered" OR "reversed"
+   ```
 
-### Direct Contradictions
+4. **Weasel words** — Notes that express uncertainty about aspects the claim treats as certain
+   - "might", "probably", "not sure", "need to verify"
+   - These indicate the claim is stronger than the evidence supports
 
-- **[[Note]]** ({date}): "{quote that contradicts the belief}"
-  **Why this matters**: {explanation}
+### Step 4: Identify Unstated Assumptions
 
-### Position Changes Over Time
+For the claim to be true, what else must be true?
 
-- {date range}: You held view X (evidence: [[notes]])
-- {date}: Something shifted (trigger: [[note]])
-- {date range}: You now hold view Y
+- Technical assumptions (about systems, tools, constraints)
+- People assumptions (about availability, skills, priorities)
+- Temporal assumptions (about deadlines, sequencing, stability)
+- External assumptions (about market, users, dependencies)
 
-### Unstated Assumptions
+Check if the vault evidence supports each assumption.
 
-1. **{Assumption}**: {Why this might not hold}
-   Evidence: {what the vault says or doesn't say}
+### Step 5: Assess Evidence Strength
 
-2. **{Assumption}**: {Why this might not hold}
+Rate the overall case:
 
-### Missing Perspectives
+| Strength | Meaning |
+|----------|---------|
+| Strong | Multiple independent sources agree, no contradictions |
+| Moderate | Some support, minor gaps or contradictions |
+| Weak | Limited evidence, significant contradictions or gaps |
+| Unknown | Vault doesn't have enough information |
 
-- {Whose view is absent and why it matters}
+### Step 6: Present the Challenge
 
-## Verdict
-
-{2-3 paragraphs — fair assessment of how the belief holds up. Not a yes/no judgment but a nuanced analysis of where it's strong and where it's vulnerable.}
-
-## Questions to Sit With
+```markdown
+# Challenge: "{claim}"
 
-1. {A question the user should consider — not rhetorical, genuinely open}
-2. {Another question}
-3. {Another question}
+## The Claim
+{Restate the claim clearly}
 
-## Suggested Actions
+## What Rests On This
+{What assumptions does the claim depend on?}
+- {Assumption 1} — {supported / unsupported / contradicted}
+- {Assumption 2} — {supported / unsupported / contradicted}
 
-- {Concrete action if the belief needs revision}
-- {Action to gather missing evidence}
-- {Notes to re-read with fresh eyes}
-```
+## Supporting Evidence
+- [[Note A]]: "{quote}"
+- [[Note B]]: "{quote}"
 
-### Step 6: Save (Optional)
+## Challenging Evidence
+- ⚠ [[Note C]]: "{quote that contradicts or weakens the claim}"
+- ⚠ [[Note D]]: "{quote showing uncertainty or alternative view}"
 
-At the end of your challenge, ask:
+## Position Changes Over Time
+- {date}: [[Note E]] said X
+- {date}: [[Note F]] said Y (contradicts X)
 
-> "Would you like me to save this as a note?"
+## Weak Points
+1. **{Weak point}**: {why it's weak, which note shows it}
+2. **{Weak point}**: {why it's weak, which note shows it}
 
-If the user confirms, save with frontmatter:
+## Overall Assessment
+**Evidence strength**: {Strong / Moderate / Weak / Unknown}
 
-```yaml
----
-title: "Challenge: {Belief}"
-note_type: literature
-type: analysis
-date: <today>
-tags: [challenge, critical-thinking]
----
+{2-3 sentence summary: what the vault evidence suggests about this claim, what's uncertain, and what would strengthen or weaken the case further}
 ```
 
-Use `obsidian create` to save. Ask the user where they'd like it saved.
-
 ## Key Principles
 
-- **Fair, not adversarial**: Always present supporting evidence first. The goal is clear thinking, not winning an argument.
-- **Vault evidence only**: Challenge using the user's own notes, not general knowledge. "Research says X" is not valid here — "Your note from March says X" is.
-- **Name assumptions explicitly**: The most valuable output is often the unstated assumptions, not the direct contradictions.
-- **Questions over conclusions**: End with questions, not verdicts. The user decides what to do with the analysis.
-- **Respect the "strong" result**: If a belief holds up well, say so clearly. Not every challenge needs to find problems.
+- **Respectful opposition.** The goal is to strengthen thinking, not to tear it down. Frame challenges as "here's what to consider" not "you're wrong."
+- **Evidence only.** Every challenge must cite specific vault notes. Don't invent external counterarguments.
+- **Surface uncertainty.** If the vault shows doubt or hesitation about aspects the claim treats as certain, highlight this gap.
+- **Obsidian is first workbench.** All note operations go through Obsidian CLI.