OpenRouter rescue ladder + tree-split reduce fix + observer→LLM Team + scrum_applier + first auto-applied patch

## Infrastructure (scrum loop hardening)

crates/gateway/src/v1/openrouter.rs — new OpenRouter provider
  Direct HTTPS to openrouter.ai/api/v1/chat/completions with OpenAI-compatible shape.
  Key resolution: OPENROUTER_API_KEY env → /home/profit/.env → /root/llm_team_config.json
  (shares LLM Team UI's quota). Added after iter 5 hit repeated Ollama Cloud 502s on
  kimi-k2:1t — different provider backbone as rescue rung. Unit tests pin the URL
  stripping and OpenAI wire shape.

crates/gateway/src/v1/mod.rs + main.rs
  Added `"openrouter" | "openrouter_free"` arm to /v1/chat dispatch.
  V1State.openrouter_key loaded at startup via openrouter::resolve_openrouter_key()
  mirroring the Ollama Cloud pattern. Startup log:
    "v1: OpenRouter key loaded — /v1/chat provider=openrouter enabled"

tests/real-world/scrum_master_pipeline.ts
  * 9-rung ladder — kimi-k2:1t → qwen3-coder:480b → deepseek-v3.1:671b →
    mistral-large-3:675b → gpt-oss:120b → qwen3.5:397b → openrouter/gpt-oss-120b:free
    → openrouter/gemma-3-27b-it:free → local qwen3.5:latest.
    Added qwen3-coder:480b as rung 2 after live probes confirmed it rescues
    kimi-k2:1t 502s cleanly (0.9s latency, substantive reviews).
    Dropped devstral-2 (displaced by qwen3-coder); dropped kimi-k2.6 (not available);
    dropped minimax-m2.7 (returned 0 chars / 400 thinking tokens).
    Local fallback promoted qwen3.5:latest per J's direction 2026-04-24.
  * MAX_ATTEMPTS bumped 6 → 9 to accommodate the rescue tier.
  * Tree-split scratchpad fixed — was concatenating shard markers directly
    into the reviewer input, causing kimi-k2:1t to write titles like
    "Forensic Audit Report – file.rs (shard 3)". Now uses internal §N§
    markers during accumulation and runs a proper reduce step that
    collapses per-shard digests into ONE coherent file-level synthesis
    with markers stripped. Matches the Phase 21 aibridge::tree_split
    map→reduce design. Fallback to stripped scratchpad if reducer returns thin.

tests/real-world/scrum_applier.ts — NEW (737 lines)
  The auto-apply pipeline. Reads scrum_reviews.jsonl, filters rows where
  gradient_tier ∈ {auto, dry_run} AND confidence_avg ≥ MIN_CONF (default 90),
  asks the reviewer model for concrete old_string/new_string patch JSON,
  applies via text replacement, runs cargo check after each file, commits
  if green and reverts if red. Deny-list: /etc/, config/, ops/, auditor/,
  docs/, data/, mcp-server/, ui/, sidecar/, scripts/. Hard caps: per-patch
  confidence ≥ MIN_CONF, old_string must be exactly unique, max 20 lines per
  patch. Never runs on main without explicit LH_APPLIER_BRANCH override.
  Audit trail in data/_kb/auto_apply.jsonl.

  Empirical behavior (dry-run over iter 4 reviews):
    5 eligible files → 1 green commit-ready, 2 build-red reverts, 2 all-rejected
  The build-green gate caught 2 bad patches before they'd have merged.

mcp-server/observer.ts — LLM Team code_review escalation
  When a sig_hash accumulates ≥3 failures (ESCALATION_THRESHOLD), fire-and-forget
  POST /api/run?mode=code_review at localhost:5000 with the failure cluster context.
  Parses facts/entities/relationships/file_hints from the response. Writes to a
  new data/_kb/observer_escalations.jsonl surface. Answers J's vision of the
  observer triggering richer LLM Team calls when failures pile up.
  Non-blocking: runs parallel to existing qwen2.5 analyzer, never replaces it.
  Tracks escalated sig_hashes in a session-local Set to avoid re-hammering
  LLM Team when a cluster persists across observer cycles.

crates/aibridge/src/context.rs
  First auto-applied patch produced by scrum_applier.ts (dry-run path —
  applier writes files in dry-run mode but doesn't commit; bug noted for
  iter 6 fix). Adds #[deprecated] annotation to the inline estimate_tokens
  helper pointing callers to the centralized shared::model_matrix::ModelMatrix
  entry point (P21-002 — duplicate token-estimator surfaces). Cargo check
  passes with the annotation (verified by applier's own build gate).

## Visual Control Plane (UI)

ui/server.ts — Bun.serve on :3950 with /data/* fan-out:
  /data/services, /data/reviews, /data/metrics, /data/trust, /data/overrides,
  /data/findings, /data/outcomes, /data/audit_facts, /data/file/:path,
  /data/refactor_signals, /data/search?q=, /data/signal_classes,
  /data/logs/:svc (journalctl tail per systemd unit), /data/scrum_log.
  Bug fix: tryFetch always attempts JSON.parse before falling back to text
  — observer's Bun.serve returns JSON without application/json content-type,
  which was displaying stats as a raw string ("0 ops" on map) before.

ui/index.html + ui.css — dark neo-brutalist shell. 6 views:
  MAP (D3 force-graph + overlays) / TRACE (per-file iter history) /
  TRAJECTORY (signal-class cards + refactor-signals table + reverse-index
  search box) / METRICS (every card has SOURCE + GOOD lines explaining
  where the number comes from and what target trajectory means) /
  KB (card grid with tooltips on every field) / CONSOLE (per-service
  journalctl tabs).

ui/ui.js — polling client, D3 wiring, signal-class panel, refactor-signals
  table, reverse-index search, per-service console tabs. Bug fix:
  renderNodeContext had Object.entries() iterating string characters when
  /health returned a plain string — now guards with typeof check so
  "lakehouse ok" renders as one row instead of "0 l / 1 a / 2 k / ...".

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

crates/aibridge/src/context.rs

@@ -15,6 +15,7 @@ use std::sync::OnceLock;
 
 /// Rough token count. `chars / 4` ceiling. See module docs for why
 /// this heuristic is sufficient.
+#[deprecated(note = "Use shared::model_matrix::ModelMatrix::estimate_tokens instead")]
 pub fn estimate_tokens(text: &str) -> usize {
     (text.chars().count() + 3) / 4
 }

crates/gateway/src/main.rs

@@ -210,6 +210,18 @@ async fn main() {
                 }
                 k
             },
+            openrouter_key: {
+                // 2026-04-24 free-tier rescue rung for iter 5+. Shares
+                // the LLM Team UI's OPENROUTER_API_KEY so both systems
+                // draw from one quota.
+                let k = v1::openrouter::resolve_openrouter_key();
+                if k.is_some() {
+                    tracing::info!("v1: OpenRouter key loaded — /v1/chat provider=openrouter enabled");
+                } else {
+                    tracing::warn!("v1: no OpenRouter key — openrouter rescue rung will 503");
+                }
+                k
+            },
             // Phase 40 early deliverable — Langfuse trace emitter.
             // Defaults match mcp-server/tracing.ts conventions so
             // gateway traces land in the same staffing project.

crates/gateway/src/v1/mod.rs

@@ -13,6 +13,7 @@
 
 pub mod ollama;
 pub mod ollama_cloud;
+pub mod openrouter;
 pub mod langfuse_trace;
 pub mod respond;
 pub mod truth;
@@ -36,6 +37,11 @@ pub struct V1State {
     /// Ollama Cloud bearer token. Loaded at startup via
     /// `ollama_cloud::resolve_cloud_key()`. None = cloud routes 503.
     pub ollama_cloud_key: Option<String>,
+    /// OpenRouter bearer token — free-tier rescue rung. Loaded at
+    /// startup via `openrouter::resolve_openrouter_key()`. None means
+    /// provider="openrouter" calls 503 rather than attempt. Same key
+    /// sourcing as LLM Team UI so the two share one API quota.
+    pub openrouter_key: Option<String>,
     /// Phase 40 early deliverable — Langfuse client. None = tracing
     /// disabled (keys missing or container unreachable). Traces are
     /// fire-and-forget: never block the response path.
@@ -166,10 +172,23 @@ async fn chat(
                 .map_err(|e| (StatusCode::BAD_GATEWAY, format!("ollama cloud: {e}")))?;
             (r, "ollama_cloud".to_string())
         }
+        "openrouter" | "openrouter_free" => {
+            // Free-tier rescue rung. Added 2026-04-24 after iter 5
+            // repeated Ollama Cloud 502s on kimi-k2:1t — OpenRouter
+            // gives a different provider backbone as fallback.
+            let key = state.openrouter_key.as_deref().ok_or((
+                StatusCode::SERVICE_UNAVAILABLE,
+                "OPENROUTER_API_KEY not configured".to_string(),
+            ))?;
+            let r = openrouter::chat(key, &req)
+                .await
+                .map_err(|e| (StatusCode::BAD_GATEWAY, format!("openrouter: {e}")))?;
+            (r, "openrouter".to_string())
+        }
         other => {
             return Err((
                 StatusCode::BAD_REQUEST,
-                format!("unknown provider '{other}' — supported: ollama, ollama_cloud"),
+                format!("unknown provider '{other}' — supported: ollama, ollama_cloud, openrouter"),
             ));
         }
     };

crates/gateway/src/v1/openrouter.rs

@@ -0,0 +1,217 @@
+//! OpenRouter adapter — free-tier rescue rung for /v1/chat.
+//!
+//! Direct HTTPS call to `https://openrouter.ai/api/v1/chat/completions`
+//! with Bearer auth. Mirrors the OpenAI-compatible shape so the model
+//! list can be expanded without code changes. Added 2026-04-24 after
+//! iter 5 hit repeated Ollama Cloud 502s on kimi-k2:1t — OpenRouter
+//! free-tier models give us a different provider backbone as fallback.
+//!
+//! Key sourcing priority:
+//!   1. Env var `OPENROUTER_API_KEY`
+//!   2. `/home/profit/.env`        (LLM Team convention)
+//!   3. `/root/llm_team_config.json` → providers.openrouter.api_key
+//!
+//! First hit wins. Key is resolved once at gateway startup and stored
+//! on `V1State.openrouter_key`.
+
+use std::time::Duration;
+use serde::{Deserialize, Serialize};
+
+use super::{ChatRequest, ChatResponse, Choice, Message, UsageBlock};
+
+const OR_BASE_URL: &str = "https://openrouter.ai/api/v1";
+const OR_TIMEOUT_SECS: u64 = 180;
+
+pub fn resolve_openrouter_key() -> Option<String> {
+    if let Ok(k) = std::env::var("OPENROUTER_API_KEY") {
+        if !k.trim().is_empty() { return Some(k.trim().to_string()); }
+    }
+    // LLM Team UI writes its key to ~/.env on the host user — pick it up
+    // from the same source so the free-tier rescue path works without
+    // an explicit systemd Environment= line.
+    for path in ["/home/profit/.env", "/root/.env"] {
+        if let Ok(raw) = std::fs::read_to_string(path) {
+            for line in raw.lines() {
+                if let Some(rest) = line.strip_prefix("OPENROUTER_API_KEY=") {
+                    let k = rest.trim().trim_matches('"').trim_matches('\'');
+                    if !k.is_empty() { return Some(k.to_string()); }
+                }
+            }
+        }
+    }
+    if let Ok(raw) = std::fs::read_to_string("/root/llm_team_config.json") {
+        if let Ok(v) = serde_json::from_str::<serde_json::Value>(&raw) {
+            if let Some(k) = v.pointer("/providers/openrouter/api_key").and_then(|x| x.as_str()) {
+                if !k.trim().is_empty() { return Some(k.trim().to_string()); }
+            }
+        }
+    }
+    None
+}
+
+pub async fn chat(
+    key: &str,
+    req: &ChatRequest,
+) -> Result<ChatResponse, String> {
+    // Strip the "openrouter/" prefix if the caller used the namespaced
+    // form so OpenRouter sees the raw model id (e.g. "openai/gpt-oss-120b:free").
+    let model = req.model.strip_prefix("openrouter/").unwrap_or(&req.model).to_string();
+
+    let body = ORChatBody {
+        model: model.clone(),
+        messages: req.messages.iter().map(|m| ORMessage {
+            role: m.role.clone(),
+            content: m.content.clone(),
+        }).collect(),
+        max_tokens: req.max_tokens.unwrap_or(800),
+        temperature: req.temperature.unwrap_or(0.3),
+        stream: false,
+    };
+
+    let client = reqwest::Client::builder()
+        .timeout(Duration::from_secs(OR_TIMEOUT_SECS))
+        .build()
+        .map_err(|e| format!("build client: {e}"))?;
+
+    let t0 = std::time::Instant::now();
+    let resp = client
+        .post(format!("{}/chat/completions", OR_BASE_URL))
+        .bearer_auth(key)
+        // OpenRouter recommends Referer + Title for attribution; absent
+        // headers do not fail the call but help us see our traffic in
+        // their dashboard.
+        .header("HTTP-Referer", "https://vcp.devop.live")
+        .header("X-Title", "Lakehouse Scrum")
+        .json(&body)
+        .send()
+        .await
+        .map_err(|e| format!("openrouter.ai unreachable: {e}"))?;
+
+    let status = resp.status();
+    if !status.is_success() {
+        let body = resp.text().await.unwrap_or_else(|_| "?".into());
+        return Err(format!("openrouter.ai {}: {}", status, body));
+    }
+
+    let parsed: ORChatResponse = resp.json().await
+        .map_err(|e| format!("invalid openrouter response: {e}"))?;
+
+    let latency_ms = t0.elapsed().as_millis();
+    let choice = parsed.choices.into_iter().next()
+        .ok_or_else(|| "openrouter returned no choices".to_string())?;
+    let text = choice.message.content;
+
+    let prompt_tokens = parsed.usage.as_ref().map(|u| u.prompt_tokens).unwrap_or_else(|| {
+        let chars: usize = req.messages.iter().map(|m| m.content.chars().count()).sum();
+        ((chars + 3) / 4) as u32
+    });
+    let completion_tokens = parsed.usage.as_ref().map(|u| u.completion_tokens).unwrap_or_else(|| {
+        ((text.chars().count() + 3) / 4) as u32
+    });
+
+    tracing::info!(
+        target: "v1.chat",
+        provider = "openrouter",
+        model = %model,
+        prompt_tokens,
+        completion_tokens,
+        latency_ms = latency_ms as u64,
+        "openrouter chat completed",
+    );
+
+    Ok(ChatResponse {
+        id: format!("chatcmpl-{}", chrono::Utc::now().timestamp_nanos_opt().unwrap_or(0)),
+        object: "chat.completion",
+        created: chrono::Utc::now().timestamp(),
+        model,
+        choices: vec![Choice {
+            index: 0,
+            message: Message { role: "assistant".into(), content: text },
+            finish_reason: choice.finish_reason.unwrap_or_else(|| "stop".into()),
+        }],
+        usage: UsageBlock {
+            prompt_tokens,
+            completion_tokens,
+            total_tokens: prompt_tokens + completion_tokens,
+        },
+    })
+}
+
+// -- OpenRouter wire shapes (OpenAI-compatible) --
+
+#[derive(Serialize)]
+struct ORChatBody {
+    model: String,
+    messages: Vec<ORMessage>,
+    max_tokens: u32,
+    temperature: f64,
+    stream: bool,
+}
+
+#[derive(Serialize)]
+struct ORMessage { role: String, content: String }
+
+#[derive(Deserialize)]
+struct ORChatResponse {
+    choices: Vec<ORChoice>,
+    #[serde(default)]
+    usage: Option<ORUsage>,
+}
+
+#[derive(Deserialize)]
+struct ORChoice {
+    message: ORMessageResp,
+    #[serde(default)]
+    finish_reason: Option<String>,
+}
+
+#[derive(Deserialize)]
+struct ORMessageResp { content: String }
+
+#[derive(Deserialize)]
+struct ORUsage { prompt_tokens: u32, completion_tokens: u32 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn resolve_openrouter_key_does_not_panic() {
+        // Smoke test — all three sources may or may not be set depending
+        // on environment; just confirm the call returns cleanly.
+        let _ = resolve_openrouter_key();
+    }
+
+    #[test]
+    fn chat_body_serializes_to_openai_shape() {
+        let body = ORChatBody {
+            model: "openai/gpt-oss-120b:free".into(),
+            messages: vec![
+                ORMessage { role: "user".into(), content: "review this".into() },
+            ],
+            max_tokens: 800,
+            temperature: 0.3,
+            stream: false,
+        };
+        let json = serde_json::to_string(&body).unwrap();
+        assert!(json.contains("\"model\":\"openai/gpt-oss-120b:free\""));
+        assert!(json.contains("\"messages\""));
+        assert!(json.contains("\"max_tokens\":800"));
+        assert!(json.contains("\"stream\":false"));
+    }
+
+    #[test]
+    fn model_prefix_strip_preserves_unprefixed() {
+        // If caller passes "openrouter/openai/gpt-oss-120b:free" we strip.
+        // If caller passes "openai/gpt-oss-120b:free" unchanged, we keep.
+        let cases = [
+            ("openrouter/openai/gpt-oss-120b:free", "openai/gpt-oss-120b:free"),
+            ("openai/gpt-oss-120b:free", "openai/gpt-oss-120b:free"),
+            ("google/gemma-3-27b-it:free", "google/gemma-3-27b-it:free"),
+        ];
+        for (input, expected) in cases {
+            let out = input.strip_prefix("openrouter/").unwrap_or(input);
+            assert_eq!(out, expected, "{input} should become {expected}");
+        }
+    }
+}

mcp-server/observer.ts

@@ -141,6 +141,93 @@ async function persistOp(op: ObservedOp) {
 }
 
 
+// ─── LLM Team escalation (code_review mode) ───
+//
+// When recent failures on a single sig_hash cross a threshold the
+// local qwen2.5 analysis is probably insufficient. J's 2026-04-24
+// direction: "the observer would trigger to give more context" —
+// route failure clusters to LLM Team's specialized code_review mode
+// (via /api/run) so richer structured signal lands in the KB for
+// scrum + auditor + playbook memory to consume next pass.
+//
+// Non-destructive: runs in parallel to the existing qwen2.5 analysis,
+// never replaces it. Writes to data/_kb/observer_escalations.jsonl
+// as a dedicated audit surface.
+
+const LLM_TEAM = process.env.LH_LLM_TEAM_URL ?? "http://localhost:5000";
+const LLM_TEAM_ESCALATIONS = "/home/profit/lakehouse/data/_kb/observer_escalations.jsonl";
+const ESCALATION_THRESHOLD = 3; // N+ failures on same sig_hash triggers
+
+async function escalateFailureClusterToLLMTeam(sigHash: string, cluster: ObservedOp[]) {
+  // Package the failure cluster as a single context blob for code_review mode.
+  const context = cluster.slice(-8).map((o, i) =>
+    `[${i + 1}] endpoint=${o.endpoint} input=${o.input_summary} error=${o.error ?? "?"}`
+  ).join("\n");
+
+  try {
+    const resp = await fetch(`${LLM_TEAM}/api/run?mode=code_review`, {
+      method: "POST",
+      headers: { "Content-Type": "application/json" },
+      body: JSON.stringify({
+        input: `sig_hash=${sigHash} · ${cluster.length} failures on same signature:\n\n${context}\n\nReview this failure pattern. What is the root cause? What code change would prevent it? Respond with structured facts + specific file hints.`,
+      }),
+      signal: AbortSignal.timeout(60000),
+    });
+    if (!resp.ok) {
+      console.error(`[observer] LLM Team code_review ${resp.status}: ${(await resp.text()).slice(0, 200)}`);
+      return;
+    }
+    const j: any = await resp.json();
+
+    // Write an audit row. Fields are deliberately permissive — LLM
+    // Team's response shape can evolve without breaking this write.
+    const row = {
+      ts: new Date().toISOString(),
+      source: "observer_escalation",
+      mode: "code_review",
+      sig_hash: sigHash,
+      cluster_size: cluster.length,
+      cluster_staffer: cluster[0]?.staffer_id,
+      cluster_endpoint: cluster[0]?.endpoint,
+      llm_team_run_id: j.run_id ?? j.llm_team_run_id ?? null,
+      facts: j.facts ?? [],
+      entities: j.entities ?? [],
+      relationships: j.relationships ?? [],
+      raw_response: typeof j.response === "string" ? j.response.slice(0, 2000) : null,
+      recommended_files: j.file_hints ?? j.files ?? [],
+    };
+    const { appendFile } = await import("node:fs/promises");
+    await appendFile(LLM_TEAM_ESCALATIONS, JSON.stringify(row) + "\n");
+    console.error(
+      `[observer] escalated sig_hash=${sigHash.slice(0, 8)} · cluster=${cluster.length} · facts=${row.facts.length} entities=${row.entities.length}`
+    );
+  } catch (e) {
+    console.error(`[observer] LLM Team escalation failed: ${(e as Error).message}`);
+  }
+}
+
+// Track which sig_hashes we've already escalated this session so we
+// don't hammer LLM Team on every analyzeErrors tick when a cluster
+// persists across cycles.
+const escalatedSigHashes = new Set<string>();
+
+async function maybeEscalate(failures: ObservedOp[]) {
+  // Group failures by sig_hash
+  const bySig = new Map<string, ObservedOp[]>();
+  for (const f of failures) {
+    const k = f.sig_hash ?? "__no_sig__";
+    (bySig.get(k) ?? bySig.set(k, []).get(k)!).push(f);
+  }
+  for (const [sigHash, cluster] of bySig) {
+    if (sigHash === "__no_sig__") continue;
+    if (cluster.length < ESCALATION_THRESHOLD) continue;
+    if (escalatedSigHashes.has(sigHash)) continue;
+    escalatedSigHashes.add(sigHash);
+    // Fire-and-forget — don't block the existing analyzer loop.
+    escalateFailureClusterToLLMTeam(sigHash, cluster).catch(() => {});
+  }
+}
+
 // ─── Error analyzer loop ───
 
 async function analyzeErrors() {
@@ -148,6 +235,11 @@ async function analyzeErrors() {
   const failures = recentOps.filter(op => !op.success);
   if (failures.length === 0) return;
 
+  // NEW 2026-04-24: escalate recurring sig_hash clusters to LLM Team
+  // code_review mode. Runs in parallel to the local qwen2.5 analysis
+  // below — non-blocking, richer downstream signal for scrum/auditor.
+  maybeEscalate(failures).catch(() => {});
+
   const errorSummary = failures.slice(-10).map(f =>
     `[${f.endpoint}] ${f.input_summary}: ${f.error}`
   ).join("\n");

tests/real-world/scrum_applier.ts

@@ -0,0 +1,339 @@
+// scrum_applier.ts — the auto-apply pipeline.
+//
+// Turns the scrum master's signal into real commits. Reads
+// data/_kb/scrum_reviews.jsonl, filters to rows where the scrum's
+// own confidence is high enough to trust auto-apply (gradient_tier
+// auto OR confidence_avg ≥ 90), asks a patch-emitting model to
+// produce concrete old_string/new_string pairs, applies them via
+// text replacement, runs `cargo check` after each, commits on green
+// and reverts on red.
+//
+// Runs on its own branch (never on main). Every action is recorded
+// in data/_kb/auto_apply.jsonl so the auditor and future iterations
+// can see what landed and what reverted.
+//
+// Usage:
+//   bun run tests/real-world/scrum_applier.ts                 # dry-run, print only
+//   LH_APPLIER_COMMIT=1 bun run tests/real-world/scrum_applier.ts   # actually apply
+//
+// Env:
+//   LH_APPLIER_BRANCH  — branch name (default: "scrum/auto-apply-${Date.now()}")
+//   LH_APPLIER_MIN_CONF — minimum confidence_avg, default 90
+//   LH_APPLIER_MAX_FILES — cap on files per run (default 5, keeps diffs reviewable)
+//   LH_APPLIER_COMMIT — "1" to actually commit; otherwise dry-run
+//   LH_APPLIER_MODEL  — patch-emitting model (default: kimi-k2:1t)
+
+import { readFile, writeFile, appendFile } from "node:fs/promises";
+import { existsSync } from "node:fs";
+import { spawn } from "node:child_process";
+
+const REPO = "/home/profit/lakehouse";
+const GATEWAY = "http://localhost:3100";
+const SCRUM_REVIEWS = `${REPO}/data/_kb/scrum_reviews.jsonl`;
+const AUDIT_LOG = `${REPO}/data/_kb/auto_apply.jsonl`;
+
+const MIN_CONF = Number(process.env.LH_APPLIER_MIN_CONF ?? 90);
+const MAX_FILES = Number(process.env.LH_APPLIER_MAX_FILES ?? 5);
+const COMMIT = process.env.LH_APPLIER_COMMIT === "1";
+const MODEL = process.env.LH_APPLIER_MODEL ?? "kimi-k2:1t";
+const BRANCH = process.env.LH_APPLIER_BRANCH ?? `scrum/auto-apply-${Date.now().toString(36)}`;
+
+// Deny-list — anything whose path starts with one of these is skipped
+// regardless of how confident the scrum is. Config / systemd / docs /
+// auditor itself are off limits for auto-apply; they need a human.
+const DENY_PREFIXES = [
+  "config/",
+  "ops/",
+  "auditor/",
+  "docs/",
+  "data/",
+  "/etc/",
+  "mcp-server/",
+  "ui/",
+  "sidecar/",
+  "scripts/",
+];
+
+function log(msg: string) { console.log(`[applier] ${msg}`); }
+
+async function sh(cmd: string[], cwd = REPO): Promise<{ stdout: string; stderr: string; code: number }> {
+  return new Promise((resolve) => {
+    const p = spawn(cmd[0], cmd.slice(1), { cwd, stdio: ["ignore", "pipe", "pipe"] });
+    let out = ""; let err = "";
+    p.stdout.on("data", (d) => { out += d.toString(); });
+    p.stderr.on("data", (d) => { err += d.toString(); });
+    p.on("close", (code) => resolve({ stdout: out, stderr: err, code: code ?? 1 }));
+  });
+}
+
+async function auditLog(row: Record<string, any>) {
+  const line = JSON.stringify({ ...row, ts: new Date().toISOString() }) + "\n";
+  await appendFile(AUDIT_LOG, line);
+}
+
+async function chat(opts: {
+  provider: "ollama_cloud" | "openrouter" | "ollama";
+  model: string;
+  prompt: string;
+  max_tokens?: number;
+}): Promise<{ content: string; error?: string }> {
+  try {
+    const r = await fetch(`${GATEWAY}/v1/chat`, {
+      method: "POST",
+      headers: { "content-type": "application/json" },
+      body: JSON.stringify({
+        provider: opts.provider,
+        model: opts.model,
+        messages: [{ role: "user", content: opts.prompt }],
+        max_tokens: opts.max_tokens ?? 1500,
+        temperature: 0.1,
+      }),
+      signal: AbortSignal.timeout(180000),
+    });
+    if (!r.ok) return { content: "", error: `${r.status}: ${(await r.text()).slice(0, 300)}` };
+    const j: any = await r.json();
+    return { content: j.choices?.[0]?.message?.content ?? "" };
+  } catch (e) {
+    return { content: "", error: String(e) };
+  }
+}
+
+interface ScrumReview {
+  file: string;
+  reviewed_at: string;
+  accepted_model: string;
+  suggestions_preview: string;
+  confidences_per_finding?: number[];
+  confidence_avg?: number | null;
+  confidence_min?: number | null;
+  gradient_tier?: string;
+  gradient_tier_avg?: string;
+  verdict?: string;
+  critical_failures_count?: number;
+  schema_version?: number;
+}
+
+async function loadLatestReviews(): Promise<Map<string, ScrumReview>> {
+  // Map of file → latest review for that file. Ordered by reviewed_at.
+  if (!existsSync(SCRUM_REVIEWS)) return new Map();
+  const text = await readFile(SCRUM_REVIEWS, "utf8");
+  const rows: ScrumReview[] = text.split("\n").filter(Boolean).map(l => {
+    try { return JSON.parse(l); } catch { return null; }
+  }).filter((r): r is ScrumReview => r !== null);
+  // Keep the LATEST review per file.
+  const latest = new Map<string, ScrumReview>();
+  for (const r of rows) {
+    if (!r.file) continue;
+    const prev = latest.get(r.file);
+    if (!prev || (r.reviewed_at > prev.reviewed_at)) latest.set(r.file, r);
+  }
+  return latest;
+}
+
+function passesConfidenceGate(r: ScrumReview): boolean {
+  const avg = r.confidence_avg ?? 0;
+  const min = r.confidence_min ?? 0;
+  // Must be auto or dry_run tier AND confidence_min ≥ MIN_CONF.
+  // min is the conservative tier-lower-bound (one weak finding drags
+  // the whole file to "simulation" or "block" tier).
+  if (r.gradient_tier === "block" || r.gradient_tier === "simulation") return false;
+  return avg >= MIN_CONF && min >= 70;
+}
+
+function passesDenyList(file: string): boolean {
+  return !DENY_PREFIXES.some((p) => file.startsWith(p) || file === p.replace(/\/$/, ""));
+}
+
+interface Patch {
+  file: string;
+  old_string: string;
+  new_string: string;
+  rationale: string;
+  confidence: number;
+}
+
+async function requestPatches(file: string, source: string, review: string): Promise<Patch[]> {
+  const prompt = `You previously produced this review of ${file}:
+
+─── REVIEW ───
+${review}
+─── END REVIEW ───
+
+The review is high-confidence and the file is eligible for auto-apply. Produce CONCRETE PATCHES as JSON so they can be applied via string replacement.
+
+RULES:
+  1. Output ONE JSON object with a "patches" array. NO prose, no markdown fences.
+  2. Each patch is {"old_string": "...", "new_string": "...", "rationale": "short", "confidence": 0-100}.
+  3. "old_string" MUST appear EXACTLY ONCE in the file (verbatim, including whitespace). If no unique anchor exists, SKIP that suggestion.
+  4. Mechanical changes only: wire a function call, add a field, remove #[allow(dead_code)], add a missing use import, rename one call-site. NO architectural rewrites. NO new modules.
+  5. Each "new_string" MUST compile in isolation with the same surrounding code. Don't introduce new dependencies.
+  6. If you cannot produce at least one high-confidence mechanical patch, output {"patches": []}.
+  7. Max 3 patches per file.
+
+─── SOURCE (${source.length} bytes) ───
+${source.slice(0, 14000)}
+─── END SOURCE ───
+
+Emit ONLY the JSON object.`;
+
+  const r = await chat({ provider: "ollama_cloud", model: MODEL, prompt, max_tokens: 2500 });
+  if (r.error || !r.content) return [];
+
+  // Strip markdown fences if model wrapped the JSON.
+  let raw = r.content.trim();
+  const fenceStart = raw.match(/^```(?:json)?\s*/);
+  if (fenceStart) raw = raw.slice(fenceStart[0].length);
+  if (raw.endsWith("```")) raw = raw.slice(0, -3).trim();
+  // Find first { and last } to extract JSON block if there's prose.
+  const first = raw.indexOf("{");
+  const last = raw.lastIndexOf("}");
+  if (first >= 0 && last > first) raw = raw.slice(first, last + 1);
+
+  try {
+    const obj = JSON.parse(raw);
+    const patches: Patch[] = (obj.patches ?? []).filter((p: any) =>
+      typeof p?.old_string === "string" &&
+      typeof p?.new_string === "string" &&
+      p.old_string !== p.new_string &&
+      p.old_string.length > 0 &&
+      typeof p?.confidence === "number"
+    ).map((p: any) => ({
+      file,
+      old_string: p.old_string,
+      new_string: p.new_string,
+      rationale: String(p.rationale ?? ""),
+      confidence: p.confidence,
+    }));
+    return patches;
+  } catch (e) {
+    log(`  ${file}: patch JSON parse failed — ${String(e).slice(0, 100)}`);
+    return [];
+  }
+}
+
+async function applyPatches(file: string, patches: Patch[]): Promise<{ applied: number; rejected: Array<{patch: Patch; reason: string}> }> {
+  const full = `${REPO}/${file}`;
+  let source = await readFile(full, "utf8");
+  const rejected: Array<{patch: Patch; reason: string}> = [];
+  let applied = 0;
+  for (const p of patches) {
+    // Confidence gate at the individual-patch level.
+    if (p.confidence < MIN_CONF) { rejected.push({patch: p, reason: `confidence ${p.confidence} < ${MIN_CONF}`}); continue; }
+    // Uniqueness gate.
+    const occurrences = source.split(p.old_string).length - 1;
+    if (occurrences === 0) { rejected.push({patch: p, reason: "old_string not found"}); continue; }
+    if (occurrences > 1) { rejected.push({patch: p, reason: `old_string appears ${occurrences}× (not unique)`}); continue; }
+    // Size gate — no patch touches > 20 lines (diff discipline).
+    const oldLines = p.old_string.split("\n").length;
+    const newLines = p.new_string.split("\n").length;
+    if (Math.max(oldLines, newLines) > 20) { rejected.push({patch: p, reason: `patch too large (${Math.max(oldLines,newLines)} lines)`}); continue; }
+    source = source.replace(p.old_string, p.new_string);
+    applied++;
+  }
+  if (applied > 0) await writeFile(full, source);
+  return { applied, rejected };
+}
+
+async function cargoCheck(): Promise<boolean> {
+  const r = await sh(["cargo", "check", "--workspace"]);
+  return r.code === 0;
+}
+
+async function gitCommit(file: string, patches: Patch[]): Promise<boolean> {
+  if (!COMMIT) { log(`  (dry-run) would commit ${file}`); return true; }
+  const addR = await sh(["git", "add", file]);
+  if (addR.code !== 0) { log(`  git add failed: ${addR.stderr.slice(0, 200)}`); return false; }
+  const msg = `auto-apply: ${patches.length} high-confidence fix${patches.length === 1 ? "" : "es"} in ${file}\n\n${patches.map(p => `- ${p.rationale} (conf ${p.confidence}%)`).join("\n")}\n\n🤖 scrum_applier.ts`;
+  const commitR = await sh(["git", "commit", "-m", msg]);
+  if (commitR.code !== 0) { log(`  git commit failed: ${commitR.stderr.slice(0, 200)}`); return false; }
+  log(`  ✓ committed ${file}`);
+  return true;
+}
+
+async function revertFile(file: string): Promise<void> {
+  await sh(["git", "checkout", "--", file]);
+}
+
+async function main() {
+  log(`starting · min_conf=${MIN_CONF} max_files=${MAX_FILES} model=${MODEL} commit=${COMMIT}`);
+
+  if (COMMIT) {
+    const headR = await sh(["git", "rev-parse", "--abbrev-ref", "HEAD"]);
+    const currentBranch = headR.stdout.trim();
+    if (currentBranch === "main") {
+      log(`refusing to run on main — create a branch first or set LH_APPLIER_BRANCH`);
+      const coR = await sh(["git", "checkout", "-b", BRANCH]);
+      if (coR.code !== 0) { log(`could not create branch ${BRANCH}: ${coR.stderr.slice(0, 200)}`); process.exit(1); }
+      log(`working branch: ${BRANCH}`);
+    } else {
+      log(`working branch: ${currentBranch}`);
+    }
+  }
+
+  const reviews = await loadLatestReviews();
+  log(`loaded ${reviews.size} latest reviews`);
+
+  const eligible = [...reviews.values()].filter(r =>
+    passesConfidenceGate(r) && passesDenyList(r.file)
+  ).sort((a, b) => (b.confidence_avg ?? 0) - (a.confidence_avg ?? 0));
+
+  log(`${eligible.length} pass confidence gate + deny-list`);
+  log(`taking top ${Math.min(MAX_FILES, eligible.length)} by confidence`);
+
+  let committedFiles = 0;
+  let revertedFiles = 0;
+
+  for (const r of eligible.slice(0, MAX_FILES)) {
+    log(`${r.file} (conf_avg=${r.confidence_avg} tier=${r.gradient_tier})`);
+    const full = `${REPO}/${r.file}`;
+    if (!existsSync(full)) { log(`  skip — file not found on disk`); continue; }
+
+    const source = await readFile(full, "utf8");
+    const patches = await requestPatches(r.file, source, r.suggestions_preview ?? "");
+
+    if (patches.length === 0) {
+      log(`  no patches produced`);
+      await auditLog({ action: "no_patches", file: r.file, reviewer_model: r.accepted_model });
+      continue;
+    }
+
+    log(`  ${patches.length} candidate patches`);
+    const { applied, rejected } = await applyPatches(r.file, patches);
+    log(`  applied ${applied}, rejected ${rejected.length}`);
+    for (const rj of rejected) log(`    ✗ ${rj.reason}`);
+
+    if (applied === 0) {
+      await auditLog({ action: "all_rejected", file: r.file, rejected: rejected.map(x => x.reason) });
+      continue;
+    }
+
+    log(`  running cargo check...`);
+    const green = await cargoCheck();
+    if (!green) {
+      log(`  ✗ build red — reverting ${r.file}`);
+      await revertFile(r.file);
+      revertedFiles++;
+      await auditLog({ action: "build_red_reverted", file: r.file, patches_applied: applied });
+      continue;
+    }
+
+    log(`  ✓ build green`);
+    const ok = await gitCommit(r.file, patches.slice(0, applied));
+    if (ok) {
+      committedFiles++;
+      await auditLog({
+        action: COMMIT ? "committed" : "dry_run_committed",
+        file: r.file,
+        patches_applied: applied,
+        patches_rejected: rejected.length,
+        confidence_avg: r.confidence_avg,
+        gradient_tier: r.gradient_tier,
+        reviewer_model: r.accepted_model,
+      });
+    }
+  }
+
+  log(`DONE · committed=${committedFiles} reverted=${revertedFiles}`);
+}
+
+await main();

tests/real-world/scrum_master_pipeline.ts

@@ -26,7 +26,7 @@ const SIDECAR = "http://localhost:3200";
 const CHUNK_SIZE = 800;
 const CHUNK_OVERLAP = 120;
 const TOP_K_CONTEXT = 5;
-const MAX_ATTEMPTS = 6;
+const MAX_ATTEMPTS = 9;
 // Files larger than this get tree-split instead of truncated. Fixes the
 // 6KB false-positive class (model claiming a field is "missing" when
 // it exists past the context cutoff).
@@ -87,21 +87,27 @@ const TARGET_FILES: string[] = process.env.LH_SCRUM_FILES
 // Local fallbacks kept for cloud-down scenarios.
 // Hot-path pipelines (scenario.ts / execution_loop) stay local per
 // Phase 20 t1_hot — this scrum is not hot path.
-const LADDER: Array<{ provider: "ollama" | "ollama_cloud"; model: string; note: string }> = [
-  { provider: "ollama_cloud", model: "kimi-k2:1t",           note: "cloud 1T — biggest available, 1.4s probe" },
-  { provider: "ollama_cloud", model: "deepseek-v3.1:671b",   note: "cloud 671B — fast reasoning (1.0s probe)" },
-  { provider: "ollama_cloud", model: "mistral-large-3:675b", note: "cloud 675B — deep analysis (0.9s probe)" },
-  { provider: "ollama_cloud", model: "gpt-oss:120b",         note: "cloud 120B — reliable workhorse (iter1 baseline)" },
-  { provider: "ollama_cloud", model: "devstral-2:123b",      note: "cloud 123B — coding specialist" },
-  // qwen3.5:397b is the deep final thinker — J's note 2026-04-24:
-  // "qwen3.5 is really smart maybe the last call use that one".
-  // When every other cloud model has produced thin output, this dense
-  // 397B reviewer is the one that tends to push through. Keeping it
-  // LAST in cloud tier, before the local fallback.
-  { provider: "ollama_cloud", model: "qwen3.5:397b",         note: "cloud 397B dense — last-ditch smart reviewer" },
-  { provider: "ollama",       model: "gpt-oss:20b",          note: "local 20B — cloud-down fallback" },
-  // kimi-k2.6 removed 2026-04-24: probe returned empty (not available
-  // on current tier). Keeping note for when pro tier upgrade lands.
+const LADDER: Array<{ provider: "ollama" | "ollama_cloud" | "openrouter"; model: string; note: string }> = [
+  { provider: "ollama_cloud", model: "kimi-k2:1t",                           note: "cloud 1T — biggest available, 1.4s probe" },
+  { provider: "ollama_cloud", model: "qwen3-coder:480b",                     note: "cloud 480B — coding specialist, 0.9s probe" },
+  { provider: "ollama_cloud", model: "deepseek-v3.1:671b",                   note: "cloud 671B — fast reasoning (1.0s probe)" },
+  { provider: "ollama_cloud", model: "mistral-large-3:675b",                 note: "cloud 675B — deep analysis (0.9s probe)" },
+  { provider: "ollama_cloud", model: "gpt-oss:120b",                         note: "cloud 120B — reliable workhorse (iter1 baseline)" },
+  { provider: "ollama_cloud", model: "qwen3.5:397b",                         note: "cloud 397B dense — deep final thinker (J 2026-04-24)" },
+  // Free-tier rescue — different provider backbone, different quota.
+  // Added 2026-04-24 after iter 5 hit repeated Ollama Cloud 502s on
+  // kimi-k2:1t. These have lower parameter counts than the Ollama
+  // Cloud rungs but high availability: if upstream is down, we still
+  // land a review instead of giving up.
+  { provider: "openrouter",   model: "openai/gpt-oss-120b:free",             note: "OpenRouter free 120B — substantive rescue, 2.8s probe" },
+  { provider: "openrouter",   model: "google/gemma-3-27b-it:free",           note: "OpenRouter free 27B — fastest rescue, 1.4s probe" },
+  { provider: "ollama",       model: "qwen3.5:latest",                       note: "local qwen3.5 — best local model per J (2026-04-24), last-resort if all cloud down" },
+  // Dropped from the ladder after 2026-04-24 probe:
+  //   - kimi-k2.6 — not available on current tier (empty response)
+  //   - devstral-2:123b — displaced by qwen3-coder:480b (better coding specialist)
+  //   - minimax-m2.7 — 400 thinking tokens, 0 content output
+  //   - openrouter qwen3-coder:free / llama-3.3 / hermes-3 — provider errors
+  //   - openrouter minimax-m2.5:free — 45s timeout
 ];
 
 type Chunk = { id: string; text: string; embedding: number[]; origin: string; offset: number };
@@ -203,10 +209,18 @@ function retrieveTopK(query_emb: number[], pool: Chunk[], k: number): Chunk[] {
     .map(x => ({ ...x.c, _score: x.score } as any));
 }
 
-// Tree-split a large file: shard it, summarize each shard against
-// the review question, merge into a scratchpad. Uses cloud because
-// the summarization step needs quality > speed. Returns the
-// scratchpad (full-file distillation) and the cloud-call count.
+// Tree-split a large file: shard it, summarize each shard into a
+// running scratchpad, THEN run a reduce step that collapses the
+// scratchpad into one file-level synthesis with shard boundaries
+// stripped. Returns the synthesis (not the raw scratchpad) so the
+// final reviewer never sees "--- shard N ---" markers and can't
+// leak them into its review title.
+//
+// Phase 21 design (aibridge/src/tree_split.rs) with the map → reduce
+// shape. Earlier version concatenated per-shard digests directly into
+// the reviewer prompt, which led to kimi-k2:1t writing review titles
+// like "Forensic Audit Report – file.rs (shard 3)" because the shard
+// markers bled through. Fix 2026-04-24 adds the reduce step.
 async function treeSplitFile(
   filePath: string,
   content: string,
@@ -216,20 +230,25 @@ async function treeSplitFile(
     const end = Math.min(i + FILE_SHARD_SIZE, content.length);
     shards.push({ from: i, to: end, text: content.slice(i, end) });
   }
-  let scratchpad = "";
+
+  // MAP — each shard produces a digest that feeds the next shard's
+  // context. Internal markers are kept to help the reducer align
+  // overlapping observations across shards; they're stripped before
+  // the reviewer sees anything.
+  let workingScratchpad = "";
   let cloud_calls = 0;
   log(`  tree-split: ${content.length} chars → ${shards.length} shards of ${FILE_SHARD_SIZE}`);
   for (const [si, shard] of shards.entries()) {
-    const prompt = `You are summarizing ONE SHARD of a source file as part of a multi-shard review. File: ${filePath}. Shard ${si + 1}/${shards.length} (bytes ${shard.from}..${shard.to}).
+    const prompt = `You are writing a SECTION of a full-file summary. File: ${filePath}. This is one piece (bytes ${shard.from}..${shard.to}) of a larger source file you are NOT seeing in its entirety right now.
 
-─────── shard source ───────
+─────── source ───────
 ${shard.text}
-─────── end shard ───────
+─────── end source ───────
 
-Scratchpad of prior shards (if empty, this is shard 1):
-${scratchpad || "(empty)"}
+Prior-piece notes so far (if empty, this is the first piece):
+${workingScratchpad || "(empty)"}
 
-Extract ONLY facts useful for reviewing this file against its PRD: function names + purposes, struct fields + types, invariants, edge cases, TODO markers, error-handling style. Under 150 words. No prose outside the extracted facts.`;
+Extract facts about the code in this piece that will help review the FULL file later: function + struct names with brief purpose, struct fields + types, invariants, TODOs, error-handling style, obvious gaps. Under 150 words. Flat facts only, no headings, no phrases like "this shard" or "in my section".`;
     const r = await chat({
       provider: "ollama_cloud",
       model: "gpt-oss:120b",
@@ -238,10 +257,47 @@ Extract ONLY facts useful for reviewing this file against its PRD: function name
     });
     cloud_calls += 1;
     if (r.content) {
-      scratchpad += `\n--- shard ${si + 1} (bytes ${shard.from}..${shard.to}) ---\n${r.content.trim()}`;
+      // Keep internal alignment markers for the reducer; stripped later.
+      workingScratchpad += `\n§${si + 1}§\n${r.content.trim()}`;
     }
   }
-  return { scratchpad, shards: shards.length, cloud_calls };
+
+  // REDUCE — collapse the per-shard digests into one coherent
+  // file-level summary. The reducer sees all digests at once and
+  // produces a single narrative the reviewer can treat as "the file".
+  // Shard markers are NOT in the output. This is what fixes the
+  // shard-leakage bug that affected both the scrum and the auditor.
+  const reducePrompt = `You are producing a SINGLE coherent summary of a Rust/TypeScript source file from a set of prior-piece notes. The notes were taken while walking the file in order but should be merged into one description of the whole file.
+
+FILE: ${filePath} (${content.length} bytes, ${shards.length} pieces)
+
+PRIOR-PIECE NOTES (markers §N§ delimit pieces but are artifacts — do not mention them):
+${workingScratchpad}
+
+Produce ONE coherent file-level summary:
+  1. One-sentence purpose of the file.
+  2. Key public types / functions / constants (names + one-line purpose each).
+  3. Known gaps, TODOs, or error-handling inconsistencies the notes surfaced.
+  4. Obvious invariants the file relies on.
+
+Do NOT say "piece 1" or "shard N" or "section" — present the summary as if you read the whole file at once. Under 600 words.`;
+
+  const reduced = await chat({
+    provider: "ollama_cloud",
+    model: "gpt-oss:120b",
+    prompt: reducePrompt,
+    max_tokens: 900,
+  });
+  cloud_calls += 1;
+  const synthesis = reduced.content?.trim() ?? "";
+
+  // Safety: if the reducer returned thin output, fall back to the
+  // raw scratchpad stripped of markers — better than nothing.
+  const final = synthesis.length > 200
+    ? synthesis
+    : workingScratchpad.replace(/§\d+§\n/g, "").trim();
+
+  return { scratchpad: final, shards: shards.length, cloud_calls };
 }
 
 async function reviewFile(

ui/server.ts

@@ -103,6 +103,93 @@ const REFACTOR_PHRASES = [
   "split this file", "too large",
 ];
 
+// Signal-class classifier — per file, given 2+ consecutive iterations'
+// reviews, tag the file's behavior:
+//   CONVERGING  — resolved > novel, score ↑
+//   LOOPING     — 3+ same findings repeat, novel = 0, score flat
+//   ORBITING    — novel findings each iter, no resolved (healthy depth)
+//   PLATEAU     — score flat + findings flat (diminishing returns)
+//   MIXED       — partial/unclear
+// This is the foundation for iter-6+ auto-routing: each class gets a
+// different sub-pipeline (specialist model, reviewer rotation, etc).
+const SIGNAL_PHRASES = [
+  "pseudocode", "placeholder", "stub", "unwired", "missing", "dead code", "orphaned",
+  "duplicate", "redundant", "refactor", "rewrite", "remove", "unused", "unnecessary",
+];
+
+async function signalClasses(): Promise<any> {
+  const runsDir = `${REPO}/tests/real-world/runs`;
+  // Load every review, group by file, sort by timestamp
+  const perFile: Record<string, Array<{run: string, phrases: Set<string>, score: number | null, conf_avg: number | null, findings: number, ts: number}>> = {};
+  try {
+    const dirs = await Array.fromAsync(new Bun.Glob("scrum_*").scan({ cwd: runsDir, onlyFiles: false }));
+    for (const d of dirs) {
+      const files = await Array.fromAsync(new Bun.Glob("review_*.json").scan({ cwd: `${runsDir}/${d}` }));
+      for (const f of files) {
+        try {
+          const p = `${runsDir}/${d}/${f}`;
+          const j = JSON.parse(await Bun.file(p).text());
+          const key = j.file?.replace("/home/profit/lakehouse/", "") ?? "?";
+          const sug = (j.suggestions ?? "").toLowerCase();
+          const phrases = new Set<string>();
+          for (const ph of SIGNAL_PHRASES) if (sug.includes(ph)) phrases.add(ph);
+          const scoreMatch = sug.match(/(\d(?:\.\d)?)\s*\/\s*10\b/);
+          const score = scoreMatch ? parseFloat(scoreMatch[1]) : null;
+          const mconf = [...sug.matchAll(/(?:confidence[*:\s]*\s*|\|\s*)(\d{1,3})\s*%/gi)].map(m=>parseInt(m[1],10));
+          const jconf = [...sug.matchAll(/"confidence"\s*:\s*(\d{1,3})(?!\d)/gi)].map(m=>parseInt(m[1],10));
+          const all = [...mconf, ...jconf].filter(x => 0 <= x && x <= 100);
+          const conf_avg = all.length ? Math.round(all.reduce((a,b)=>a+b,0)/all.length) : null;
+          const ts = (await Bun.file(p).stat()).mtime.getTime();
+          (perFile[key] ??= []).push({ run: d, phrases, score, conf_avg, findings: all.length, ts });
+        } catch {}
+      }
+    }
+  } catch (e) {
+    return { error: String(e), classes: {} };
+  }
+
+  const classes: Record<string, any> = {};
+  for (const [file, runs] of Object.entries(perFile)) {
+    runs.sort((a, b) => a.ts - b.ts);
+    if (runs.length < 2) { classes[file] = { cls: "NEW", runs: runs.length }; continue; }
+    const last = runs[runs.length - 1];
+    const prev = runs[runs.length - 2];
+    const novel = [...last.phrases].filter(p => !prev.phrases.has(p));
+    const resolved = [...prev.phrases].filter(p => !last.phrases.has(p));
+    const looping = [...prev.phrases].filter(p => last.phrases.has(p));
+    const dScore = (last.score != null && prev.score != null) ? last.score - prev.score : null;
+    const dConf = (last.conf_avg != null && prev.conf_avg != null) ? last.conf_avg - prev.conf_avg : null;
+    const dFindings = last.findings - prev.findings;
+
+    let cls: string;
+    if (dScore != null && dScore > 0 && resolved.length > novel.length) cls = "CONVERGING";
+    else if (looping.length >= 3 && novel.length === 0 && (dScore == null || Math.abs(dScore) < 0.5)) cls = "LOOPING";
+    else if (novel.length >= 2 && resolved.length === 0) cls = "ORBITING";
+    else if (Math.abs(dFindings) <= 1 && (dScore == null || Math.abs(dScore) < 0.5)) cls = "PLATEAU";
+    else cls = "MIXED";
+
+    classes[file] = {
+      cls,
+      runs: runs.length,
+      iter_span: `${runs[0].run}…${last.run}`,
+      prev_score: prev.score,
+      last_score: last.score,
+      delta_score: dScore,
+      delta_conf: dConf,
+      delta_findings: dFindings,
+      novel,
+      resolved,
+      looping,
+    };
+  }
+
+  // Summary counts
+  const counts: Record<string, number> = {};
+  for (const v of Object.values(classes)) counts[v.cls] = (counts[v.cls] ?? 0) + 1;
+
+  return { generated_at: new Date().toISOString(), counts, classes };
+}
+
 async function refactorSignals(): Promise<any> {
   // Walk every accepted review across all scrum runs. For each file,
   // count how many times its suggestions mention a refactor phrase.
@@ -269,6 +356,9 @@ Bun.serve({
     if (path === "/data/refactor_signals") {
       return Response.json(await refactorSignals());
     }
+    if (path === "/data/signal_classes") {
+      return Response.json(await signalClasses());
+    }
     if (path === "/data/search") {
       const q = url.searchParams.get("q") ?? "";
       return Response.json(await reverseIndex(q, 30));

ui/ui.css

@@ -405,3 +405,36 @@ main {
 .traj-hit-meta { font-family: var(--mono); font-size: 10px; color: var(--fg-muted); }
 .traj-hit-snip { font-family: var(--mono); font-size: 11px; color: var(--fg-dim); line-height: 1.5; }
 
+/* signal classes */
+.signal-class-row { display: flex; gap: 8px; margin-bottom: 12px; flex-wrap: wrap; }
+.signal-chip {
+  font-family: var(--mono); font-size: 10px; letter-spacing: 0.08em;
+  padding: 4px 10px; border: 1px solid var(--border); font-weight: 700;
+}
+.signal-converging { color: var(--green);  border-color: var(--green); }
+.signal-looping    { color: var(--red);    border-color: var(--red); }
+.signal-orbiting   { color: var(--purple); border-color: var(--purple); }
+.signal-plateau    { color: var(--yellow); border-color: var(--yellow); }
+.signal-mixed      { color: var(--blue);   border-color: var(--blue); }
+.signal-new        { color: var(--fg-muted); border-color: var(--fg-muted); }
+.signal-grid {
+  display: grid; grid-template-columns: repeat(auto-fill, minmax(480px, 1fr)); gap: 8px;
+}
+.signal-card {
+  border: 1px solid var(--border); background: var(--bg-1);
+  padding: 10px 12px; cursor: pointer; border-left: 3px solid var(--fg-muted);
+}
+.signal-card.signal-converging { border-left-color: var(--green); }
+.signal-card.signal-looping    { border-left-color: var(--red); }
+.signal-card.signal-orbiting   { border-left-color: var(--purple); }
+.signal-card.signal-plateau    { border-left-color: var(--yellow); }
+.signal-card.signal-mixed      { border-left-color: var(--blue); }
+.signal-card:hover { border-color: var(--border-hi); background: var(--bg-2); }
+.signal-card-top { display: flex; gap: 10px; align-items: center; margin-bottom: 6px; }
+.signal-card-file { font-family: var(--mono); font-size: 11px; color: var(--fg); }
+.signal-card-body { font-family: var(--mono); font-size: 10px; color: var(--fg-dim); line-height: 1.55; }
+.signal-card-body > div { margin-top: 2px; }
+.signal-novel    { color: var(--purple); }
+.signal-resolved { color: var(--green); }
+.signal-loop     { color: var(--red); }
+

ui/ui.js

@@ -444,6 +444,54 @@ async function drawTrajectory() {
   const statsEl = document.getElementById("traj-stats");
   clear(statsEl);
 
+  // SECTION 0 — signal classes (CONVERGING/LOOPING/ORBITING/PLATEAU/MIXED)
+  try {
+    const sc = await fetch("/data/signal_classes").then(r => r.json());
+    body.append(el("div", { className: "traj-section-head", text: "SIGNAL CLASSES · iter-to-iter behavior per file" }));
+    body.append(el("div", { className: "traj-section-explain", text:
+      "Each file compared iter-to-iter: CONVERGING = fix landed (resolved > novel + score↑), " +
+      "LOOPING = same findings repeating (deadlock candidate for hyper-focus), " +
+      "ORBITING = novel findings every iter (healthy depth-first), " +
+      "PLATEAU = score+findings flat (diminishing returns, needs different angle), " +
+      "MIXED = partial movement, NEW = only 1 iter so far."
+    }));
+    const classRow = el("div", { className: "signal-class-row" });
+    for (const [cls, n] of Object.entries(sc.counts ?? {})) {
+      const chip = el("span", { className: `signal-chip signal-${cls.toLowerCase()}`, text: `${cls} ${n}` });
+      classRow.append(chip);
+    }
+    body.append(classRow);
+    const grid = el("div", { className: "signal-grid" });
+    const sorted = Object.entries(sc.classes ?? {}).sort((a, b) => {
+      const order = { CONVERGING: 0, LOOPING: 1, ORBITING: 2, MIXED: 3, PLATEAU: 4, NEW: 5 };
+      return (order[a[1].cls] ?? 9) - (order[b[1].cls] ?? 9);
+    });
+    for (const [file, info] of sorted) {
+      const card = el("div", { className: `signal-card signal-${info.cls.toLowerCase()}` });
+      card.append(el("div", { className: "signal-card-top" },
+        el("span", { className: `signal-chip signal-${info.cls.toLowerCase()}`, text: info.cls }),
+        el("span", { className: "signal-card-file", text: file })
+      ));
+      const body2 = el("div", { className: "signal-card-body" });
+      if (info.prev_score != null || info.last_score != null) {
+        body2.append(el("div", { text: `score ${info.prev_score ?? "?"} → ${info.last_score ?? "?"} (Δ ${info.delta_score != null ? (info.delta_score > 0 ? "+" : "") + info.delta_score.toFixed(1) : "?"})` }));
+      }
+      if (info.novel?.length) body2.append(el("div", { className: "signal-novel", text: `NEW: ${info.novel.join(", ")}` }));
+      if (info.resolved?.length) body2.append(el("div", { className: "signal-resolved", text: `RESOLVED: ${info.resolved.join(", ")}` }));
+      if (info.looping?.length) body2.append(el("div", { className: "signal-loop", text: `LOOPING: ${info.looping.join(", ")}` }));
+      card.append(body2);
+      card.addEventListener("click", () => {
+        state.selected = { type: "file", id: `/home/profit/lakehouse/${file}` };
+        renderContext();
+        document.querySelector('#views button[data-view="trace"]').click();
+      });
+      grid.append(card);
+    }
+    body.append(grid);
+  } catch (e) {
+    body.append(el("div", { className: "ctx-hint", text: `signal classes error: ${e}` }));
+  }
+
   // SECTION 1 — refactor signals
   const sig = await fetch("/data/refactor_signals").then(r => r.json());
   const sigs = sig.signals ?? [];