Auto: entities/internet-finance/hyperspace.md | 1 file changed, 25 insertions(+)

CLAUDE.md

@@ -4,6 +4,19 @@
 
 If you're exploring this repo with Claude Code, you're talking to a **collective knowledge base** maintained by 6 AI domain specialists. ~400 claims across 14 knowledge areas, all linked, all traceable from evidence through claims through beliefs to public positions.
 
+### Contributor Recognition
+
+Before orientation, check if this person is a known contributor. Look up their identity (X handle, name, or however they introduce themselves) against `contributors.json` or the attribution data in the knowledge base.
+
+**If they're a known contributor:** Skip orientation. Load their contributor card and engage at their tier level:
+
+- **Contributor tier:** Reference their history. "You challenged Rio's claim about Dutch auctions last month — that challenge is still standing after 2 counter-attempts. What are you working on now?" Then load the relevant agent and engage.
+- **Veteran tier:** Peer engagement. Reference shared history, ask for their take on open questions, invite them to specific gaps in the KB where their expertise is most valuable. "We have a gap in futarchy redistribution evidence — you've been the strongest voice on this. Want to help us close it?"
+
+The agents remember contributors and treat them accordingly. This is earned, not granted — it comes from visible contribution history in the knowledge base.
+
+**If they're unknown or new:** Run the visitor orientation below.
+
 ### Orientation (run this on first visit)
 
 Don't present a menu. Start a short conversation to figure out who this person is and what they care about.

domains/internet-finance/access-friction-functions-as-a-natural-conviction-filter-in-token-launches-because-process-difficulty-selects-for-genuine-believers-while-price-friction-selects-for-wealthy-speculators.md

@@ -0,0 +1,62 @@
+---
+type: claim
+domain: internet-finance
+description: "Purchase friction (technical barriers, bridging requirements, OTC-only access) filters for conviction via time/effort cost rather than capital cost, producing a qualitatively different holder base than price-based mechanisms like Dutch auctions — ordinals OTC and Hyperliquid pre-CEX are the strongest empirical cases"
+confidence: experimental
+source: "Rio via @CryptoTomYT friction-is-bullish thesis; ordinals OTC market data; Hyperliquid pre-CEX trading data"
+created: 2026-03-16
+secondary_domains:
+  - mechanisms
+depends_on:
+  - "early-conviction pricing is an unsolved mechanism design problem because systems that reward early believers attract extractive speculators while systems that prevent speculation penalize genuine supporters"
+  - "token launches are hybrid-value auctions where common-value price discovery and private-value community alignment require different mechanisms because auction theory optimized for one degrades the other"
+challenged_by:
+  - "Survivorship bias: we only observe the friction-gated assets that succeeded. The majority of friction-gated assets (ordinals that went to zero, early DeFi protocols) produced terrible outcomes."
+  - "Access friction may simply correlate with early timing, and early timing in bull markets produces outsized returns regardless of friction mechanism."
+---
+
+# Access friction functions as a natural conviction filter in token launches because process difficulty selects for genuine believers while price friction selects for wealthy speculators
+
+The early-conviction pricing trilemma identifies three properties no single mechanism achieves simultaneously: shill-proof, community-aligned, and price-discovering. The existing analysis focuses on **price friction** — mechanisms where the cost of participation is denominated in capital (Dutch auctions, bonding curves, batch auctions). But there is a fourth variable the trilemma framework doesn't capture: **access friction**, where the cost of participation is denominated in time, effort, and technical knowledge.
+
+Access friction and price friction filter for different populations:
+
+| Friction type | Cost denominated in | Filters for | Filters against |
+|---------------|--------------------|--------------|-----------------|
+| **Price friction** (Dutch auction) | Capital | Wealthy participants with high valuation | True believers who are capital-poor but conviction-rich |
+| **Access friction** (OTC-only, bridging, technical barriers) | Time, effort, knowledge | Genuine conviction (willingness to invest effort) | Extractive speculators (effort isn't scalable like capital) |
+
+**Empirical cases:**
+
+**Ordinals OTC (2023-2024):** Early Bitcoin ordinals required running a Bitcoin node, understanding the UTXO model, and negotiating OTC deals through Discord or Telegram — no marketplaces existed. This created extreme access friction. The population that navigated this friction was overwhelmingly high-conviction Bitcoin-native holders, not extractive speculators. Outcome: 6-figure single NFT sales became common among early OTC participants. When marketplaces later reduced friction (Magic Eden, Ordinals Wallet), the speculative population arrived — and returns for new entrants declined sharply.
+
+**Hyperliquid pre-CEX (2024-2025):** HYPE was only available on Hyperliquid's native platform, requiring users to bridge assets to Arbitrum and learn a new trading interface. This access friction meant early HYPE holders had already demonstrated commitment by using the product. When CEX listings eventually removed the friction, the early cohort held positions that had appreciated to 7-8 figure values. The access friction didn't prevent price discovery — Hyperliquid's own order book provided that — but it did ensure the initial holder base was product users, not pure speculators.
+
+**Why access friction is mechanistically distinct from price friction:**
+
+1. **Effort doesn't scale like capital.** A bot can deploy $10M in a Dutch auction. A bot cannot navigate an OTC negotiation requiring trust-building over Discord. Access friction resists automation in ways that price friction does not.
+
+2. **Access friction is temporarily self-limiting.** Friction decreases as infrastructure improves (marketplaces, CEX listings, bridges). This creates a natural time window where conviction-filtered holders accumulate before the friction-free speculators arrive. Price friction is permanent by design (Dutch auctions always start high).
+
+3. **Access friction doesn't penalize true believers.** In a Dutch auction, the highest-conviction buyer pays the highest price. With access friction, the highest-conviction buyer pays the same price as others who clear the access barrier — the cost is effort, not capital. This is more community-aligned.
+
+**Where access friction fails:**
+
+- **It's not a designable mechanism.** Access friction is typically accidental (early infrastructure limitations), not intentional. Once infrastructure improves, the friction disappears. You can't keep a token permanently friction-gated without killing liquidity.
+- **Survivorship bias is severe.** We observe ordinals and Hyperliquid because they succeeded. The hundreds of friction-gated assets that went to zero are invisible in this analysis.
+- **Access friction may simply proxy for timing.** Early buyers in any bull market asset tend to outperform. The friction may be incidental to the timing, not causal.
+
+**Connection to the trilemma:** Access friction suggests a possible **fourth mechanism layer** in the layered launch architecture thesis: a time-limited access-friction phase (product-only access, no CEX listings, technical barriers) that precedes the price-discovery phase. This would let conviction-filtered holders accumulate before the broader market prices the asset. The sequence: access-friction phase → price-discovery phase → open market. Effectively what Hyperliquid did accidentally.
+
+---
+
+Relevant Notes:
+- [[early-conviction pricing is an unsolved mechanism design problem because systems that reward early believers attract extractive speculators while systems that prevent speculation penalize genuine supporters]] — the trilemma this claim extends with access-friction as a fourth variable
+- [[optimal token launch architecture is layered not monolithic because separating quality governance from price discovery from liquidity bootstrapping from community rewards lets each layer use the mechanism best suited to its objective]] — access friction as a possible additional layer
+- [[dutch-auction dynamic bonding curves solve the token launch pricing problem by combining descending price discovery with ascending supply curves eliminating the instantaneous arbitrage that has cost token deployers over 100 million dollars on Ethereum]] — price-friction approach that access friction complements
+- [[futardio-cult-raised-11-4-million-in-one-day-through-futarchy-governed-meme-coin-launch]] — did Futardio Cult succeed partly because futard.io itself had access friction? Testable hypothesis.
+- [[consumer-crypto-adoption-requires-apps-optimized-for-earning-and-belonging-not-speculation]] — tension: access friction contradicts the adoption thesis. Long-term these can't coexist — friction must be temporary.
+
+Topics:
+- [[internet finance and decision markets]]
+- [[coordination mechanisms]]

domains/internet-finance/cryptographic-stake-weighted-trust-solves-autonomous-agent-coordination-without-central-authority-because-agentrank-adapts-pagerank-to-verifiable-computational-contribution.md

@@ -0,0 +1,49 @@
+---
+type: claim
+domain: internet-finance
+description: "Hyperspace's AgentRank protocol anchors autonomous agent trust to cryptographically verified computational stake, adapting PageRank to P2P agent networks — a mechanism design for reputation in multi-agent systems where no central evaluator exists"
+confidence: speculative
+source: "Rio via @varun_mathur, Hyperspace AI; AgentRank whitepaper (March 15, 2026)"
+created: 2026-03-16
+secondary_domains:
+  - ai-alignment
+  - mechanisms
+depends_on:
+  - "speculative markets aggregate information through incentive and selection effects not wisdom of crowds"
+flagged_for:
+  - theseus
+challenged_by:
+  - "Single empirical test (333 experiments, 35 agents). Scale and adversarial robustness are untested."
+  - "Computational stake may create plutocratic dynamics where GPU-rich agents dominate rankings regardless of experiment quality."
+---
+
+# Cryptographic stake-weighted trust solves autonomous agent coordination without central authority because AgentRank adapts PageRank to verifiable computational contribution
+
+Hyperspace's AgentRank (March 2026) demonstrates a mechanism design for trust among autonomous agents in decentralized networks. The core insight: when agents operate autonomously without human supervision, trust must be anchored to something verifiable. AgentRank uses cryptographically verified computational stake — proof that an agent committed real resources to its claimed experiments.
+
+**How it works:**
+1. Agents on a P2P network run ML experiments autonomously
+2. When an agent finds an improvement, it broadcasts results via GossipSub (pub/sub protocol)
+3. Other agents verify the claimed results by checking computational proofs
+4. AgentRank scores each agent based on endorsements from other agents, weighted by the endorser's own stake and track record
+5. The resulting trust graph enables the network to distinguish high-quality experimenters from noise without any central evaluator
+
+**Empirical evidence (thin):** On March 8-9 2026, 35 agents on the Hyperspace network ran 333 unsupervised experiments training language models on astrophysics papers. H100 GPU agents discovered aggressive learning rates through brute force. CPU-only laptop agents concentrated on initialization strategies and normalization techniques. The network produced differentiated research strategies without human direction, and agents learned from each other's results in real-time.
+
+**Internet finance relevance:** AgentRank is a specific implementation of the broader mechanism design problem: how do you create incentive-compatible trust in decentralized systems? The approach mirrors prediction market mechanisms — stake your resources (capital or compute), be evaluated on outcomes, build reputation through track record. The key difference: prediction markets require human judgment to define questions and settle outcomes. AgentRank operates in domains where experiment results are objectively verifiable (did the model improve?), bypassing the oracle problem.
+
+**Open questions:**
+- Does stake-weighted trust create GPU plutocracy? If ranking is proportional to compute committed, well-resourced agents dominate regardless of insight quality.
+- How does the system handle adversarial agents that fabricate computational proofs?
+- Can this mechanism generalize beyond objectively-verifiable domains (ML experiments) to domains requiring judgment (investment decisions, governance)?
+
+---
+
+Relevant Notes:
+- [[speculative markets aggregate information through incentive and selection effects not wisdom of crowds]] — AgentRank uses similar mechanism: stake creates incentive, track record creates selection
+- [[expert staking in Living Capital uses Numerai-style bounded burns for performance and escalating dispute bonds for fraud creating accountability without deterring participation]] — parallel staking mechanism for human experts, AgentRank does the same for autonomous agents
+- [[all agents running the same model family creates correlated blind spots that adversarial review cannot catch because the evaluator shares the proposers training biases]] — Hyperspace's heterogeneous compute (H100 vs CPU) naturally creates diversity. Mechanism design insight for our own pipeline.
+
+Topics:
+- [[internet finance and decision markets]]
+- [[coordination mechanisms]]

entities/internet-finance/hyperspace.md

@@ -0,0 +1,25 @@
+---
+type: entity
+entity_type: protocol
+name: "Hyperspace"
+domain: internet-finance
+secondary_domains:
+  - ai-alignment
+description: "Distributed autonomous AI agent network with P2P experiment sharing via GossipSub and stake-weighted trust via AgentRank"
+website: "https://hyper.space"
+founded: 2024
+key_people:
+  - "Varun Mathur (CEO)"
+status: active
+created: 2026-03-16
+---
+
+# Hyperspace
+
+Distributed autonomous agent network where AI agents collaborate on ML research via peer-to-peer gossip protocol. Agents share experiment results in real-time, learn from each other, and build trust through cryptographically verified computational stake (AgentRank).
+
+Key milestone: March 8-9 2026, 35 agents ran 333 unsupervised ML experiments on astrophysics papers. Heterogeneous compute (H100 GPUs vs CPU laptops) produced differentiated research strategies without human direction.
+
+AgentRank (released March 15 2026) adapts PageRank to autonomous agents, anchoring endorsements to verifiable compute contribution.
+
+Originally an "Agentic OS" / browser platform. Pivoted to distributed autonomous research infrastructure.

inbox/archive/2026-03-16-cryptotomyt-friction-is-bullish.md

@@ -0,0 +1,24 @@
+---
+type: source
+source_type: x-post
+url: "https://x.com/CryptoTomYT"
+author: "@CryptoTomYT"
+captured_date: 2026-03-16
+status: processed
+processed_date: 2026-03-16
+processed_by: rio
+claims_extracted:
+  - "access-friction-functions-as-a-natural-conviction-filter-in-token-launches-because-process-difficulty-selects-for-genuine-believers-while-price-friction-selects-for-wealthy-speculators"
+priority: standard
+notes: "Routed by Leo from Cory's X feed. Thesis: 'The more friction it is to buy, typically the best outcomes.' Evidence cited: ordinals OTC (6-figure single NFTs requiring technical knowledge + OTC negotiation), Hyperliquid (7-8 figure positions when only accessible on own platform before CEX listings). Maps to early-conviction pricing trilemma but adds novel access-friction vs price-friction distinction."
+---
+
+# CryptoTom — Friction-is-Bullish Thesis
+
+Core claim: Purchase friction (difficulty of the buying process itself) correlates with better investment outcomes because it self-selects for genuine conviction over extractive speculation.
+
+Evidence cases:
+1. **Ordinals OTC era:** Bitcoin ordinals required technical knowledge (running a node, understanding UTXO model) + OTC negotiation (no marketplaces initially). Buyers who navigated this friction were disproportionately high-conviction holders. 6-figure single NFT outcomes.
+2. **Hyperliquid pre-CEX:** When HYPE was only available on Hyperliquid's own platform (requiring bridging to Arbitrum, learning a new UI), early buyers were self-selected for conviction. 7-8 figure positions by the time CEX listings removed the friction.
+
+Mechanism claim: access friction functions as a natural Sybil filter and conviction test. The cost of overcoming process friction is denominated in time and effort, not capital — which filters differently than price-based mechanisms.

inbox/archive/2026-03-16-varun-mathur-hyperspace-distributed-agents.md

@@ -0,0 +1,27 @@
+---
+type: source
+source_type: x-post
+url: "https://x.com/varun_mathur/status/2031004607426498574"
+author: "@varun_mathur"
+captured_date: 2026-03-16
+status: processed
+processed_date: 2026-03-16
+processed_by: rio
+claims_extracted:
+  - "cryptographic-stake-weighted-trust-solves-autonomous-agent-coordination-without-central-authority-because-agentrank-adapts-pagerank-to-verifiable-computational-contribution"
+entities_extracted:
+  - "hyperspace"
+priority: standard
+flagged_for_theseus: true
+notes: "Routed by Leo from Cory's X feed. Distributed autonomous ML research lab on Hyperspace P2P network. 35 agents ran 333 unsupervised experiments via GossipSub protocol. AgentRank adapts PageRank to autonomous agents with cryptographic stake. Primary domain is AI/multi-agent (Theseus). IF angle: economic mechanism design of AgentRank (stake-weighted trust for autonomous agents)."
+---
+
+# Varun Mathur — Hyperspace Distributed Autonomous Agents
+
+March 8-9 2026: 35 autonomous agents on Hyperspace network ran 333 unsupervised ML experiments training character-level language models on astrophysics papers.
+
+Key mechanism: GossipSub P2P protocol for experiment result sharing. When an agent finds an improvement, it broadcasts to the entire network in real-time. Agents learn from each other's experiments.
+
+AgentRank (released March 15 2026): Adapts PageRank to autonomous AI agents in decentralized networks. Anchors endorsements to cryptographically verified computational stake. Economic mechanism for trust without central authority.
+
+Cross-domain note: Hyperspace took Karpathy's single-agent autoresearch loop and distributed it across P2P network. The "Autoquant" framing from Cory's intake may refer to applying this to quantitative research — distributed quant research where agents explore strategy space collaboratively.

schemas/attribution.md

@@ -97,17 +97,13 @@ A single commit may create 10 claims. The trailer says Rio committed them. The a
 
 ## Contributor Profiles
 
-Contributor profiles are reconstructed from the knowledge base, not stored separately. To build a profile:
+Contributor profiles are reconstructed from the knowledge base, not stored separately. See `schemas/contributor.md` for the full profile schema, tier system, and agent behavior differentiation.
 
-1. **Query**: search all claim `attribution` blocks for a given `handle`
-2. **Aggregate**: count contributions by role, domain, confidence level, date
-3. **Visualize**: contribution-over-time graphic showing when and how they contributed
-
-This means:
-- No separate "contributor database" to maintain
-- Profiles are always consistent with the actual KB state
-- New contributions automatically appear in profiles
-- Attribution disputes are resolved by editing claim frontmatter
+Key points:
+- Profiles are computed from attribution data, not stored as primary data
+- Three tiers (visitor → contributor → veteran) determine how agents engage
+- Contributors earn preferential treatment: agents remember their history, reference past contributions, and engage more deeply
+- See `core/reward-mechanism.md` for how attribution feeds into Contribution Index (CI) and economic rewards
 
 ### Person Entity Bridge
 

schemas/contribution-weights.yaml

@@ -1,30 +1,36 @@
 # Contribution Weights
 #
 # Global policy for how much each contributor role counts toward weighted scores.
-# Used by the build pipeline (extract-graph-data.py) to compute weighted_score
-# in contributors.json. Updated via PR — changes here affect all contributor profiles.
+# Used by the build pipeline to compute weighted_score in contributors.json
+# and Contribution Index (CI) in reward-mechanism.md.
+# Updated via PR — changes here affect all contributor profiles.
 #
 # Weights sum to 1.0. The build pipeline multiplies each contributor's role count
 # by the corresponding weight, then sums across roles.
 #
-# Current rationale (2026-03-11):
-# - Extraction is the current bottleneck and requires the most skill (reading sources,
-#   separating signal from noise, writing prose-as-title). Highest weight.
-# - Challenge is the quality mechanism — adversarial review catches errors that
-#   self-review cannot. Second highest. This also signals that the system values
-#   intellectual honesty over agreement: challenging bad claims is rewarded more
-#   than rubber-stamping good ones.
-# - Sourcing discovers new information but is lower effort per instance.
+# Current rationale (2026-03-14, revised from Rio's mechanism design brief):
+# - Sourcer = Extractor = Challenger at 0.25 each. This signals that finding
+#   the right source with a clear rationale, turning it into a structured claim,
+#   and challenging existing claims are equally valuable acts. Equal weighting
+#   prevents agent CI domination during bootstrap (agents fill extractor role,
+#   humans fill sourcer and challenger roles).
 # - Synthesis connects claims across domains — high value but rare.
-# - Review is essential but is partially automated via the eval pipeline.
+# - Review is essential but partially automated via the eval pipeline.
 #
-# These weights WILL change as the collective matures. When challenges become
-# the bottleneck (more claims than reviewers), challenger weight should increase.
-# When synthesis becomes the primary value-add, synthesizer weight increases.
+# Review after 6 months of data. If sourcer contributions turn out to be
+# low-effort, the weight is too high. If challengers produce disproportionate
+# belief changes, the weight is too low. Weights are policy, not physics.
 
 role_weights:
-  sourcer: 0.15
-  extractor: 0.40
-  challenger: 0.20
+  sourcer: 0.25
+  extractor: 0.25
+  challenger: 0.25
   synthesizer: 0.15
   reviewer: 0.10
+
+# Contribution Index (CI) leaderboard weights
+# See core/reward-mechanism.md for full spec
+ci_weights:
+  belief_movers: 0.30
+  challenge_champions: 0.30
+  connection_finders: 0.40

schemas/contributor.md

@@ -0,0 +1,132 @@
+# Contributor Schema
+
+Contributors are people who have engaged with the knowledge base. A contributor profile is computed from attribution data across claims — not stored separately. This schema defines the profile structure and tier system.
+
+## Contributor Tiers
+
+Tiers determine how agents engage with a contributor. Tier is computed from contribution history, not self-declared.
+
+| Tier | Criteria | Agent Behavior |
+|------|----------|----------------|
+| **visitor** | No contributions. First interaction. | Orientation mode: "What are you working on?" → match to agent → surface provocative claims → invite engagement. |
+| **contributor** | ≥1 merged contribution (source, challenge, or claim) | Skip orientation. Reference their history. Engage with their specific expertise. "You challenged Rio's claim about Dutch auctions — that challenge is still standing. What are you working on now?" |
+| **veteran** | ≥10 merged contributions AND ≥1 surviving challenge or belief influence | Peer engagement. Reference shared history. Invite to specific KB gaps matching their expertise. Ask for their take on open questions. Deeper context per interaction. |
+
+**Tier transitions are automatic.** The system computes tier from contribution data. No manual promotion. No application process. Contribute, and the agents notice.
+
+## Profile Structure
+
+```yaml
+handle: "@naval"                    # primary identity (X handle preferred)
+tier: contributor                   # computed: visitor | contributor | veteran
+linked_identities:                  # other identities for the same person
+  - type: x
+    handle: "@naval"
+  - type: github
+    handle: "naval"
+  - type: email
+    handle: "naval@example.com"
+first_contribution: 2026-02-15
+latest_contribution: 2026-03-11
+
+# Role counts (from attribution frontmatter across all claims)
+roles:
+  sourcer: 12
+  extractor: 0
+  challenger: 3
+  synthesizer: 1
+  reviewer: 0
+
+# Weighted score (role_counts × contribution-weights.yaml)
+weighted_score: 5.4
+
+# CI components (from reward-mechanism.md)
+ci:
+  belief_movers: 0.0
+  challenge_champions: 2.1
+  connection_finders: 0.8
+  total: 1.22    # weighted sum per ci_weights
+
+# Domain footprint
+domains:
+  internet-finance: 8
+  grand-strategy: 5
+  ai-alignment: 3
+
+# Contribution highlights (for agent context loading)
+highlights:
+  - "Challenged futarchy redistribution claim — challenge survived 2 counter-attempts"
+  - "Sourced 5 Theia Research pieces that produced 12 claims"
+  - "Connected prediction market volume claim to AI alignment belief"
+
+# Contribution timeline (monthly granularity)
+timeline:
+  - month: "2026-02"
+    count: 3
+    domains: ["internet-finance"]
+  - month: "2026-03"
+    count: 13
+    domains: ["internet-finance", "grand-strategy"]
+```
+
+## Identity Resolution
+
+**Primary identity: X handle.** X is the most likely first intake channel (people replying to claim tweets). The X handle is the canonical contributor identity.
+
+**Linked identities:** A contributor may have multiple identities across platforms (X, GitHub, email, wallet). These link to a single profile. Identity linking happens:
+- Automatically: same X handle appears in `proposed_by` (source) and git commits
+- Manually: contributor requests linking via the website or direct engagement
+
+**Pseudonymous-first.** Contributors use handles, not legal names. A handle persists across all contributions and is the public-facing identity.
+
+## How Profiles Are Computed
+
+Profiles are **derived, not stored** as primary data. The primary data is attribution frontmatter on claims and sources.
+
+### Computation steps
+
+1. **Scan all claims** for `attribution` blocks (see `schemas/attribution.md`)
+2. **Scan all sources** for `proposed_by` field
+3. **Group by handle** — aggregate role counts, domains, dates
+4. **Apply weights** from `schemas/contribution-weights.yaml`
+5. **Compute tier** from criteria above
+6. **Generate highlights** — top 3 contributions by impact (belief changes, surviving challenges, cross-domain connections)
+
+### Build artifact
+
+The build pipeline produces `contributors.json` — a static file rebuilt on every merge to main. Agents and the website read this file. No runtime queries needed.
+
+For agent session loading, a **contributor card** (compact summary) is extracted:
+
+```
+@naval | contributor | 16 contributions across internet-finance, grand-strategy
+Highlights: challenged futarchy redistribution (survived), sourced 12 Theia claims
+Last active: 2026-03-11
+```
+
+This card is injected into the agent's context at session start. ~50 tokens per contributor — cheap enough to load for any known contributor.
+
+## Agent Context Loading
+
+When a known contributor engages:
+
+1. **Lookup:** Match their identity (X handle, email, etc.) against `contributors.json`
+2. **Load card:** Inject contributor card into agent system prompt
+3. **Adjust behavior:** Agent follows tier-appropriate engagement pattern (see tiers above)
+4. **Reference history:** Agent can cite specific contributions, surviving challenges, domain expertise
+
+When an unknown person engages:
+1. **Default to visitor tier**
+2. **Run orientation flow** (see CLAUDE.md visitor section)
+3. **After first contribution:** profile is created, tier updates to contributor on next merge
+
+## Person Entity Bridge
+
+When a contributor has enough contributions to warrant tracking as an entity, their person entity (`entities/{domain}/{handle}.md`) gains `contributor: true`. The person entity tracks public information (role, organizations, influence). The contributor profile tracks KB-specific contribution data. Both link to each other.
+
+## Governance
+
+- Profiles are computed, not editable. To change your profile, change the underlying attribution data (via PR).
+- Handle changes require updating attribution frontmatter across affected claims (PR review required).
+- Disputes about attribution are resolved through the normal PR process.
+- Contributor data is public. Contribution history is visible to all agents and users.