loki-mode 8.45.0 → 8.47.0

package/SKILL.md

@@ -3,7 +3,7 @@ name: loki-mode
 description: Autonomous spec-driven build system with a built-in trust layer. It does not call work done until it is verified (RARV-C closure loop, 8 quality gates, completion council, verified-completion evidence gate). Triggers on "Loki Mode". Takes a spec (PRD, GitHub issue, OpenAPI doc, etc.) to deployed product with minimal human intervention. Provider-agnostic. Requires --dangerously-skip-permissions flag.
 ---
 
-# Loki Mode v8.45.0
+# Loki Mode v8.47.0
 
 **You are an autonomous agent. You make decisions. You do not ask questions. You do not stop.**
 
@@ -469,4 +469,4 @@ See `CHANGELOG.md` entries [7.5.7], [7.5.8], [7.5.13] for the per-fix list and r
 
 ---
 
-**v8.45.0 | [Autonomi](https://www.autonomi.dev/) flagship product | ~410 lines core**
+**v8.47.0 | [Autonomi](https://www.autonomi.dev/) flagship product | ~410 lines core**

package/VERSION

@@ -1 +1 @@
-8.45.0
+8.47.0

package/autonomy/run.sh

@@ -10583,7 +10583,6 @@ _loki_gate_stuck() {
     local threshold="${LOKI_GATE_STUCK_THRESHOLD:-3}"
 
     [ "${LOKI_GATE_STUCK_ABORT:-1}" = "0" ] && return 1
-    [ "${count:-0}" -lt "$threshold" ] 2>/dev/null && return 1
     [ -n "$reason_file" ] && [ -f "$reason_file" ] || return 1
 
     local cur prev_file prev
@@ -10592,11 +10591,26 @@ _loki_gate_stuck() {
     cur="$(head -1 "$reason_file" 2>/dev/null)" || return 1
     [ -n "$cur" ] || return 1
 
+    # RECORD ON EVERY FAILURE, compare only at threshold.
+    #
+    # The recording used to sit behind the threshold check, so the first
+    # comparison could not happen until count == threshold+1. Replayed against
+    # the REAL FireLater artifact that motivated this feature, the abort fired
+    # at iteration 4 -- and that run ended at 3. The safety valve would have
+    # missed the exact case it was built for, by one iteration.
+    #
+    # Found only by replaying the preserved .loki/quality/mutation-findings.txt
+    # rather than trusting the unit test, which used synthetic counts and so
+    # never exercised the real arrival order.
     prev_file="${TARGET_DIR:-.}/.loki/quality/gate-stuck-${gate_name}.last"
     prev="$(cat "$prev_file" 2>/dev/null || true)"
     ( mkdir -p "$(dirname "$prev_file")" 2>/dev/null \
         && printf '%s\n' "$cur" > "$prev_file" 2>/dev/null ) || true
 
+    # Below threshold: the reason is now on record for the next comparison, but
+    # this is not yet enough evidence to stop.
+    [ "${count:-0}" -lt "$threshold" ] 2>/dev/null && return 1
+
     [ -n "$prev" ] && [ "$prev" = "$cur" ] && return 0
     return 1
 }

package/dashboard/__init__.py

@@ -7,7 +7,7 @@ Modules:
     control: Session control API (start/stop/pause/resume)
 """
 
-__version__ = "8.45.0"
+__version__ = "8.47.0"
 
 # Expose the control app for easy import
 try:

package/docs/FIRST-PASS-COMPLETION-PLAN.md

@@ -65,7 +65,11 @@ work would have touched it.
 
 Ranked by measured contribution to iteration count.
 
-### F0 -- A gate that cannot pass must abort, not iterate (highest value)
+**STATUS 2026-08-01: F0, F2 and F3 SHIPPED. F1 was already built (verified, no
+work needed). F4's mechanism exists; only its default is open, and that needs
+real-build measurement rather than a guess.**
+
+### F0 -- SHIPPED v8.45.0. A gate that cannot pass aborts instead of iterating.
 
 FireLater burned 3 iterations against a gate whose detector did not exist. The
 gate correctly fail-closed each time; nothing noticed it was failing for the
@@ -81,7 +85,25 @@ same unfixable reason.
 - Guard rails: only on a byte-identical repeated cause, never a first failure,
   and it must map to a terminal-failure exit -- never a fake green.
 
-### F1 -- Front-load the 82%: make the spec gate the first-pass gate
+### F1 -- ALREADY BUILT (verified 2026-08-01). No work needed.
+
+The plan claimed this needed extending. It does not. Verified end to end:
+
+- `spec_interrogation_class_for()` already classifies findings as
+  **ambiguous / underspecified / missing / contradictory** -- the classes the
+  20,574-session study names, not just contradictions as the plan asserted.
+- `spec_ledger_prompt_block()` renders them and `run.sh:19399` assigns the
+  result to `assumption_context`.
+- `assumption_context` is interpolated into the built prompt at four sites
+  (run.sh:19685, 19687, 19691, 19693), so high-severity spec gaps reach the
+  agent on iteration 1.
+- `LOKI_SPEC_GRILL` defaults to 1, so it runs by default.
+
+**Recorded as a finding rather than converted into a change that was not
+needed.** The 82% planning bucket already has its lever; the honest next step
+is MEASURING what it catches per run, not building a second one.
+
+### F1-original (superseded) -- front-load the 82% via the spec gate
 
 `LOKI_SPEC_GRILL` already interrogates the spec before the loop and defaults
 ON. That is the correct lever for the 82% planning bucket, and it is already
@@ -95,7 +117,7 @@ paid for.
   **-3% success, +20% cost**; human-written ones **+4%**. So this must produce
   *questions and resolutions*, never a generated context blob.
 
-### F2 -- Findings injection must not be silently optional
+### F2 -- SHIPPED v8.44.0. Findings injection can no longer degrade silently.
 
 `LOKI_INJECT_FINDINGS` defaults on, but the injection is gated on
 `command -v bun`. **Without bun, the agent is told it failed and not what to
@@ -105,7 +127,7 @@ fix** -- the exact "no error recovery" shape that is 56% of failures.
 - A silent degradation of the feedback loop is worse than a missing feature,
   because the next iteration looks like a model failure.
 
-### F3 -- Verify before the gate, not after
+### F3 -- SHIPPED v8.46.0. Iteration 1 now names the gates that will judge it.
 
 47.8% first-iteration pass rate is the industry number. The cheap deterministic
 gates (test_suite, static_analysis, lsp_diagnostics) cost ~6s combined and run
@@ -117,15 +139,22 @@ gates (test_suite, static_analysis, lsp_diagnostics) cost ~6s combined and run
   edits the offending files, re-runs" -- their headline architectural change.
   We already have the gates; we just run them too late to help pass 1.
 
-### F4 -- Iteration budget as a measured decision
+### F4 -- Iteration budget: mechanism exists, default is the open question
+
+Verified: hitting the cap already records the named terminal
+`max_iterations_reached` with exit 20 (run.sh:21433) -- it does NOT fake
+success. That half of the plan was already satisfied.
 
-Research: 1-2 iteration caps fail even when the approach was sound; 5-10 is the
-recommended range. We ship `LOKI_MAX_ITERATIONS=1000`.
+What is genuinely open is the DEFAULT. `LOKI_MAX_ITERATIONS` defaults to
+**1000**; the 5 applies only to `LOKI_AUTO_FIX` tasks. Research puts the sound
+range at 5-10 and shows 1-2 fails even when the approach was correct.
 
-- The goal is not a small cap. It is **finishing in one** and stopping honestly
-  when one is not enough.
-- Pairs with F0: a cap is a blunt instrument; a named terminal reason is a
-  diagnosis.
+**Deliberately not changed here.** F0 (v8.45.0) already stops a doomed run at
+the cause rather than the count, which is the better instrument -- a cap is
+blunt, a named stuck-gate reason is a diagnosis. Lowering the default without
+measuring how often real runs legitimately exceed 5 would trade one arbitrary
+number for another. That measurement needs real builds, so it stays a founder
+decision rather than a guess.
 
 ## 4. What we do NOT do
 

package/loki-ts/dist/loki.js

@@ -1,5 +1,5 @@
 // @bun
-var h_=Object.create;var{getPrototypeOf:v_,defineProperty:rK,getOwnPropertyNames:g_}=Object;var m_=Object.prototype.hasOwnProperty;function u_(Z){return this[Z]}var p_,d_,c_=(Z,X,Q)=>{var Y=Z!=null&&typeof Z==="object";if(Y){var J=X?p_??=new WeakMap:d_??=new WeakMap,z=J.get(Z);if(z)return z}Q=Z!=null?h_(v_(Z)):{};let K=X||!Z||!Z.__esModule?rK(Q,"default",{value:Z,enumerable:!0}):Q;for(let $ of g_(Z))if(!m_.call(K,$))rK(K,$,{get:u_.bind(Z,$),enumerable:!0});if(Y)J.set(Z,K);return K};var qQ=(Z,X)=>()=>(X||Z((X={exports:{}}).exports,X),X.exports);var l_=(Z)=>Z;function i_(Z,X){this[Z]=l_.bind(null,X)}var l0=(Z,X)=>{for(var Q in X)rK(Z,Q,{get:X[Q],enumerable:!0,configurable:!0,set:i_.bind(X,Q)})};var p=(Z,X)=>()=>(Z&&(X=Z(Z=0)),X);var e0=import.meta.require;var kO={};l0(kO,{lokiDir:()=>j0,homeLokiDir:()=>P4,findRepoRootForVersion:()=>eK,REPO_ROOT:()=>r0});import{resolve as n7,dirname as tK}from"path";import{fileURLToPath as a_}from"url";import{existsSync as GQ}from"fs";import{homedir as s_}from"os";function n_(){let Z=RO;for(let X=0;X<6;X++){if(GQ(n7(Z,"VERSION"))&&GQ(n7(Z,"autonomy/run.sh")))return Z;let Q=tK(Z);if(Q===Z)break;Z=Q}return n7(RO,"..","..","..")}function eK(Z){let X=Z;for(let Q=0;Q<6;Q++){if(GQ(n7(X,"VERSION"))&&GQ(n7(X,"autonomy/run.sh")))return X;let Y=tK(X);if(Y===X)break;X=Y}return n7(Z,"..","..","..")}function j0(){return process.env.LOKI_DIR??n7(process.cwd(),".loki")}function P4(){return n7(s_(),".loki")}var RO,r0;var G8=p(()=>{RO=tK(a_(import.meta.url));r0=n_()});import{readFileSync as o_}from"fs";import{resolve as r_,dirname as t_}from"path";import{fileURLToPath as e_}from"url";function f3(){if(h5!==null)return h5;let Z="8.45.0";if(typeof Z==="string"&&Z.length>0)return h5=Z,h5;try{let X=t_(e_(import.meta.url)),Q=eK(X);h5=o_(r_(Q,"VERSION"),"utf-8").trim()}catch{h5="unknown"}return h5}var h5=null;var HQ=p(()=>{G8()});var bO={};l0(bO,{runOrThrow:()=>Of,run:()=>E0,readStreamCapped:()=>UQ,commandVersion:()=>Af,commandExists:()=>X9,ShellError:()=>Z$,MAX_STDOUT_BYTES:()=>yO});async function UQ(Z,X=yO){let Q=Z.getReader(),Y=new TextDecoder,J="",z=0;try{while(z<X){let{done:K,value:$}=await Q.read();if(K)break;if(!$)continue;if(z+=$.byteLength,z>X){let W=$.byteLength-(z-X);J+=Y.decode($.subarray(0,W),{stream:!0});break}J+=Y.decode($,{stream:!0})}J+=Y.decode()}finally{try{await Q.cancel()}catch{}Q.releaseLock()}return J}async function E0(Z,X={}){let Q=Bun.spawn({cmd:[...Z],stdout:"pipe",stderr:"pipe",env:X.env?{...process.env,...X.env}:process.env,cwd:X.cwd}),Y,J;if(X.timeoutMs&&X.timeoutMs>0)Y=setTimeout(()=>{try{Q.kill("SIGTERM")}catch{}J=setTimeout(()=>{try{Q.kill("SIGKILL")}catch{}},2000)},X.timeoutMs);try{let[z,K,$]=await Promise.all([UQ(Q.stdout),new Response(Q.stderr).text(),Q.exited]);return{stdout:z,stderr:K,exitCode:$}}finally{if(Y)clearTimeout(Y);if(J)clearTimeout(J)}}async function Of(Z,X={}){let Q=await E0(Z,X);if(Q.exitCode!==0)throw new Z$(`command failed (${Q.exitCode}): ${Z.join(" ")}`,Q.exitCode,Q.stdout,Q.stderr);return Q}async function X9(Z){let X=Lf(Z),Q=await E0(["sh","-c",`command -v ${X}`],{timeoutMs:5000});if(Q.exitCode===0)return Q.stdout.trim()||null;return null}function Lf(Z){if(!/^[A-Za-z0-9._/-]+$/.test(Z))throw Error(`refused to shell-escape suspect token: ${Z}`);return Z}async function Af(Z,X="--version"){if(!await X9(Z))return null;let Y=await E0([Z,X],{timeoutMs:5000});if(Y.exitCode!==0)return null;return((Y.stdout||Y.stderr).split(/\r?\n/)[0]?.trim()??"")||null}var yO=16777216,Z$;var x9=p(()=>{Z$=class Z$ extends Error{message;exitCode;stdout;stderr;constructor(Z,X,Q,Y){super(Z);this.message=Z;this.exitCode=X;this.stdout=Q;this.stderr=Y;this.name="ShellError"}}});function o7(Z){return jf?"":Z}var jf,L0,k8,p0,ZV0,i0,H8,Q9,v;var S6=p(()=>{jf=(process.env.NO_COLOR??"").length>0;L0=o7("\x1B[0;31m"),k8=o7("\x1B[0;32m"),p0=o7("\x1B[1;33m"),ZV0=o7("\x1B[0;34m"),i0=o7("\x1B[0;36m"),H8=o7("\x1B[1m"),Q9=o7("\x1B[2m"),v=o7("\x1B[0m")});import{existsSync as xf}from"fs";async function E7(){if(k4!==void 0)return k4;let Z="/opt/homebrew/bin/python3.12";if(xf(Z))return k4=Z,Z;let X=await X9("python3.12");if(X)return k4=X,X;let Q=await X9("python3");return k4=Q,Q}async function Y7(Z,X={}){let Q=await E7();if(!Q)return{stdout:"",stderr:"python3 not found",exitCode:127};return E0([Q,"-c",Z],X)}var k4;var r7=p(()=>{x9()});var ZL={};l0(ZL,{runStatus:()=>Yh});import{existsSync as Y9,readFileSync as v3,readdirSync as iO,statSync as aO}from"fs";import{resolve as h8,basename as sf}from"path";import{homedir as nf}from"os";function sO(Z){let X=Math.trunc(Z);if(X>=1e6)return`${(Math.trunc(X/1e6*10)/10).toFixed(1)}M`;if(X>=1000)return`${(Math.trunc(X/1000*10)/10).toFixed(1)}K`;return String(X)}function nO(Z,X,Q){if(X===0)return null;let Y=Math.trunc(Z*100/X),J=Math.trunc(Z*NQ/X);if(J>NQ)J=NQ;let z=NQ-J,K=k8;if(Y>=80)K=L0;else if(Y>=50)K=p0;let $="=".repeat(Math.max(0,J))+" ".repeat(Math.max(0,z)),W=sO(Z),V=sO(X);return`  ${H8}${Q}${v} ${K}[${$}]${v} ${Y}% (${W} / ${V})`}async function rf(){if(await X9("jq"))return!0;return process.stdout.write(`${L0}Error: jq is required but not installed.${v}
+var h_=Object.create;var{getPrototypeOf:v_,defineProperty:rK,getOwnPropertyNames:g_}=Object;var m_=Object.prototype.hasOwnProperty;function u_(Z){return this[Z]}var p_,d_,c_=(Z,X,Q)=>{var Y=Z!=null&&typeof Z==="object";if(Y){var J=X?p_??=new WeakMap:d_??=new WeakMap,z=J.get(Z);if(z)return z}Q=Z!=null?h_(v_(Z)):{};let K=X||!Z||!Z.__esModule?rK(Q,"default",{value:Z,enumerable:!0}):Q;for(let $ of g_(Z))if(!m_.call(K,$))rK(K,$,{get:u_.bind(Z,$),enumerable:!0});if(Y)J.set(Z,K);return K};var qQ=(Z,X)=>()=>(X||Z((X={exports:{}}).exports,X),X.exports);var l_=(Z)=>Z;function i_(Z,X){this[Z]=l_.bind(null,X)}var l0=(Z,X)=>{for(var Q in X)rK(Z,Q,{get:X[Q],enumerable:!0,configurable:!0,set:i_.bind(X,Q)})};var p=(Z,X)=>()=>(Z&&(X=Z(Z=0)),X);var e0=import.meta.require;var kO={};l0(kO,{lokiDir:()=>j0,homeLokiDir:()=>P4,findRepoRootForVersion:()=>eK,REPO_ROOT:()=>r0});import{resolve as n7,dirname as tK}from"path";import{fileURLToPath as a_}from"url";import{existsSync as GQ}from"fs";import{homedir as s_}from"os";function n_(){let Z=RO;for(let X=0;X<6;X++){if(GQ(n7(Z,"VERSION"))&&GQ(n7(Z,"autonomy/run.sh")))return Z;let Q=tK(Z);if(Q===Z)break;Z=Q}return n7(RO,"..","..","..")}function eK(Z){let X=Z;for(let Q=0;Q<6;Q++){if(GQ(n7(X,"VERSION"))&&GQ(n7(X,"autonomy/run.sh")))return X;let Y=tK(X);if(Y===X)break;X=Y}return n7(Z,"..","..","..")}function j0(){return process.env.LOKI_DIR??n7(process.cwd(),".loki")}function P4(){return n7(s_(),".loki")}var RO,r0;var G8=p(()=>{RO=tK(a_(import.meta.url));r0=n_()});import{readFileSync as o_}from"fs";import{resolve as r_,dirname as t_}from"path";import{fileURLToPath as e_}from"url";function f3(){if(h5!==null)return h5;let Z="8.47.0";if(typeof Z==="string"&&Z.length>0)return h5=Z,h5;try{let X=t_(e_(import.meta.url)),Q=eK(X);h5=o_(r_(Q,"VERSION"),"utf-8").trim()}catch{h5="unknown"}return h5}var h5=null;var HQ=p(()=>{G8()});var bO={};l0(bO,{runOrThrow:()=>Of,run:()=>E0,readStreamCapped:()=>UQ,commandVersion:()=>Af,commandExists:()=>X9,ShellError:()=>Z$,MAX_STDOUT_BYTES:()=>yO});async function UQ(Z,X=yO){let Q=Z.getReader(),Y=new TextDecoder,J="",z=0;try{while(z<X){let{done:K,value:$}=await Q.read();if(K)break;if(!$)continue;if(z+=$.byteLength,z>X){let W=$.byteLength-(z-X);J+=Y.decode($.subarray(0,W),{stream:!0});break}J+=Y.decode($,{stream:!0})}J+=Y.decode()}finally{try{await Q.cancel()}catch{}Q.releaseLock()}return J}async function E0(Z,X={}){let Q=Bun.spawn({cmd:[...Z],stdout:"pipe",stderr:"pipe",env:X.env?{...process.env,...X.env}:process.env,cwd:X.cwd}),Y,J;if(X.timeoutMs&&X.timeoutMs>0)Y=setTimeout(()=>{try{Q.kill("SIGTERM")}catch{}J=setTimeout(()=>{try{Q.kill("SIGKILL")}catch{}},2000)},X.timeoutMs);try{let[z,K,$]=await Promise.all([UQ(Q.stdout),new Response(Q.stderr).text(),Q.exited]);return{stdout:z,stderr:K,exitCode:$}}finally{if(Y)clearTimeout(Y);if(J)clearTimeout(J)}}async function Of(Z,X={}){let Q=await E0(Z,X);if(Q.exitCode!==0)throw new Z$(`command failed (${Q.exitCode}): ${Z.join(" ")}`,Q.exitCode,Q.stdout,Q.stderr);return Q}async function X9(Z){let X=Lf(Z),Q=await E0(["sh","-c",`command -v ${X}`],{timeoutMs:5000});if(Q.exitCode===0)return Q.stdout.trim()||null;return null}function Lf(Z){if(!/^[A-Za-z0-9._/-]+$/.test(Z))throw Error(`refused to shell-escape suspect token: ${Z}`);return Z}async function Af(Z,X="--version"){if(!await X9(Z))return null;let Y=await E0([Z,X],{timeoutMs:5000});if(Y.exitCode!==0)return null;return((Y.stdout||Y.stderr).split(/\r?\n/)[0]?.trim()??"")||null}var yO=16777216,Z$;var x9=p(()=>{Z$=class Z$ extends Error{message;exitCode;stdout;stderr;constructor(Z,X,Q,Y){super(Z);this.message=Z;this.exitCode=X;this.stdout=Q;this.stderr=Y;this.name="ShellError"}}});function o7(Z){return jf?"":Z}var jf,L0,k8,p0,ZV0,i0,H8,Q9,v;var S6=p(()=>{jf=(process.env.NO_COLOR??"").length>0;L0=o7("\x1B[0;31m"),k8=o7("\x1B[0;32m"),p0=o7("\x1B[1;33m"),ZV0=o7("\x1B[0;34m"),i0=o7("\x1B[0;36m"),H8=o7("\x1B[1m"),Q9=o7("\x1B[2m"),v=o7("\x1B[0m")});import{existsSync as xf}from"fs";async function E7(){if(k4!==void 0)return k4;let Z="/opt/homebrew/bin/python3.12";if(xf(Z))return k4=Z,Z;let X=await X9("python3.12");if(X)return k4=X,X;let Q=await X9("python3");return k4=Q,Q}async function Y7(Z,X={}){let Q=await E7();if(!Q)return{stdout:"",stderr:"python3 not found",exitCode:127};return E0([Q,"-c",Z],X)}var k4;var r7=p(()=>{x9()});var ZL={};l0(ZL,{runStatus:()=>Yh});import{existsSync as Y9,readFileSync as v3,readdirSync as iO,statSync as aO}from"fs";import{resolve as h8,basename as sf}from"path";import{homedir as nf}from"os";function sO(Z){let X=Math.trunc(Z);if(X>=1e6)return`${(Math.trunc(X/1e6*10)/10).toFixed(1)}M`;if(X>=1000)return`${(Math.trunc(X/1000*10)/10).toFixed(1)}K`;return String(X)}function nO(Z,X,Q){if(X===0)return null;let Y=Math.trunc(Z*100/X),J=Math.trunc(Z*NQ/X);if(J>NQ)J=NQ;let z=NQ-J,K=k8;if(Y>=80)K=L0;else if(Y>=50)K=p0;let $="=".repeat(Math.max(0,J))+" ".repeat(Math.max(0,z)),W=sO(Z),V=sO(X);return`  ${H8}${Q}${v} ${K}[${$}]${v} ${Y}% (${W} / ${V})`}async function rf(){if(await X9("jq"))return!0;return process.stdout.write(`${L0}Error: jq is required but not installed.${v}
 `),process.stdout.write(`Install with:
 `),process.stdout.write(`  brew install jq    (macOS)
 `),process.stdout.write(`  apt install jq     (Debian/Ubuntu)
@@ -817,6 +817,11 @@ This precedence override is narrow. It does NOT relax any safety rule. Every saf
 2. SELF-VERIFY BY RUNNING, not by reading. Run the build and the tests yourself; for each acceptance criterion, DRIVE the actual path and observe the result (submit the form, then reload and confirm the record persisted; hit a protected route logged-out and confirm it is rejected). Fix what fails now, in this pass. Do not mark done on "looks right" or a self-claim -- observed behavior is the only proof.
 3. LOCK THE ARCHITECTURE on this pass so later edits are small and additive. Decide the data model, routes, and component structure up front and build to them; never rewrite whole files later to patch a small thing (that is the doom loop that breaks working features).
 4. DESIGN: commit to ONE named aesthetic direction up front (editorial, brutalist, luxury, retro-futuristic, soft/pastel, industrial, etc. -- chosen from the product domain) and hold it on every surface. Use real content (never lorem). AVOID the AI-slop tells that instantly read as machine-generated: NO indigo/blue-to-purple gradient (the #1 tell), NO Inter/Roboto/system-font headlines (pick a real display+body pairing), NO three-equal-rounded-cards-in-a-row skeleton, NO flat 1px gray card borders or colored left-border strips, NO untouched shadcn defaults, NO reflexive dark mode. Cap the palette at ~3 hues (60/30/10), tinted not pure #fff/#000, separate sections by whitespace then a slight background shift before any border. Aim for Linear/Stripe/Duolingo-tier taste: "this does not look AI-generated".
+5. SATISFY THE GATES THAT WILL JUDGE THIS PASS. Your work is checked by deterministic gates before it can be accepted, and in measured runs EVERY extra iteration was caused by one of these rejecting the work -- never by the model failing to finish. Clear them now, in this pass:
+   - STATIC ANALYSIS: no syntax errors, no unused/undefined symbols, no lint errors in files you touched.
+   - TEST SUITE: the existing tests must still pass. Run them; do not assume.
+   - MUTATION/MOCK INTEGRITY: tests must assert real behaviour. No tautological assertions (expect(true).toBe(true)), no test that passes whether or not the code works, no mocking the very unit under test, no inline mock data standing in for a real query.
+   - CODE REVIEW: no scope creep beyond the stated task, no dead or commented-out code, no leftover debug output, and follow the conventions already in the surrounding files.
 Deliver the finished, self-verified, genuinely-designed result in THIS pass. Additional iterations should be the exception, not the plan.
 `+Kd();hj0=process.env.LOKI_CAVEMAN_VERSION||"1.9.0"});var Gq={};l0(Gq,{writeOrchestratorState:()=>Ww,updateStatusTxt:()=>gd,updateCurrentPhase:()=>hd,saveStateForRunner:()=>xd,saveState:()=>Kw,readProviderName:()=>vd,readOrchestratorState:()=>fd,loadStateForRunner:()=>Sd,loadState:()=>$w,atomicWriteFileSync:()=>S7,__setFsForTesting:()=>Dd});import{closeSync as Md,copyFileSync as Zw,existsSync as $Z,fsyncSync as Td,mkdirSync as Wq,openSync as wd,readFileSync as oY,readdirSync as Cd,renameSync as Vq,statSync as Fd,unlinkSync as rY,writeFileSync as Xw}from"fs";import{join as F2}from"path";function Dd(Z){Qw=Z.renameSync??Vq,Yw=Z.copyFileSync??Zw,Jw=Z.writeFileSync??Xw,nY=Z.unlinkSync??rY}function J3(Z){return Z??j0()}function zw(Z){return F2(Z,"autonomy-state.json")}function qq(Z){return F2(Z,"state","orchestrator.json")}function Ed(Z){return F2(Z,"state","provider")}function Id(Z){return F2(Z,"STATUS.txt")}function tT(Z){let X=null;try{X=wd(Z,"r"),Td(X)}catch{}finally{if(X!==null)try{Md(X)}catch{}}}function S7(Z,X){if(Z.endsWith(".lock"))throw Error(`atomicWriteFileSync: target path "${Z}" ends in .lock which collides with the lockfile naming convention; rename the target`);let Q=`${Z}.tmp.${process.pid}`;G6(Z,()=>{Jw(Q,X),tT(Q);try{Qw(Q,Z),RQ(Z)}catch(Y){if(Y?.code==="EXDEV")try{Yw(Q,Z),tT(Z),RQ(Z);try{nY(Q)}catch{}return}catch(z){try{nY(Q)}catch{}throw z}try{nY(Q)}catch{}throw Y}},{timeoutMs:Rd,staleMs:Pd})}function kd(Z){let X=Z.toISOString(),Q=X.indexOf(".");return Q>=0?`${X.slice(0,Q)}Z`:X}function $q(Z){return JSON.stringify(Z)}function Kw(Z){let X=J3(Z.lokiDirOverride);Wq(X,{recursive:!0});let Q=Z.now??new Date,Y=Z.prdPath??"",J=Z.pid??process.pid,z=`{
     "retryCount": ${Z.retryCount},
@@ -1222,4 +1227,4 @@ Set LOKI_LEGACY_BASH=1 to force the bash CLI for every command.
 `),2}case"start":{let{runStart:Y}=await Promise.resolve().then(() => (b_(),y_));return Y(Q)}default:return process.stderr.write(`Unknown command: ${X}
 `),process.stderr.write(__),2}}cO();process.on("SIGINT",()=>process.exit(130));process.on("SIGTERM",()=>process.exit(143));var fW0=await _W0(Bun.argv.slice(2));process.exit(fW0);
 
-//# debugId=F6128F50CF1C0A0064756E2164756E21
+//# debugId=F46363BF25356A3464756E2164756E21

package/mcp/__init__.py

@@ -75,4 +75,4 @@ try:
 except ImportError:
     __all__ = ['mcp']
 
-__version__ = '8.45.0'
+__version__ = '8.47.0'

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "loki-mode",
   "mcpName": "io.github.asklokesh/loki-mode",
-  "version": "8.45.0",
+  "version": "8.47.0",
   "description": "Loki Mode by Autonomi. Autonomous spec-to-product system: takes a PRD, GitHub issue, OpenAPI/JSON/YAML, or one-line brief to a deployed app via the RARV-C closure loop with 8 quality gates. Provider-agnostic (Claude Code, OpenAI Codex, Cline, Aider).",
   "keywords": [
     "agent",

package/plugins/loki-mode/.claude-plugin/plugin.json

@@ -2,7 +2,7 @@
   "$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
   "name": "loki-mode",
   "displayName": "Loki Mode",
-  "version": "8.45.0",
+  "version": "8.47.0",
   "description": "Autonomous spec-to-product build system with a built-in trust layer (RARV-C closure loop, 8 quality gates, completion council). Ships Loki's spec-hardening, drift-detection, and deterministic PR verification commands plus the Loki MCP server.",
   "author": {
     "name": "Autonomi",

package/providers/claude.sh

@@ -327,6 +327,11 @@ LOKI_AUTONOMY_EOF
 2. SELF-VERIFY BY RUNNING, not by reading. Run the build and the tests yourself; for each acceptance criterion, DRIVE the actual path and observe the result (submit the form, then reload and confirm the record persisted; hit a protected route logged-out and confirm it is rejected). Fix what fails now, in this pass. Do not mark done on "looks right" or a self-claim -- observed behavior is the only proof.
 3. LOCK THE ARCHITECTURE on this pass so later edits are small and additive. Decide the data model, routes, and component structure up front and build to them; never rewrite whole files later to patch a small thing (that is the doom loop that breaks working features).
 4. DESIGN: commit to ONE named aesthetic direction up front (editorial, brutalist, luxury, retro-futuristic, soft/pastel, industrial, etc. -- chosen from the product domain) and hold it on every surface. Use real content (never lorem). AVOID the AI-slop tells that instantly read as machine-generated: NO indigo/blue-to-purple gradient (the #1 tell), NO Inter/Roboto/system-font headlines (pick a real display+body pairing), NO three-equal-rounded-cards-in-a-row skeleton, NO flat 1px gray card borders or colored left-border strips, NO untouched shadcn defaults, NO reflexive dark mode. Cap the palette at ~3 hues (60/30/10), tinted not pure #fff/#000, separate sections by whitespace then a slight background shift before any border. Aim for Linear/Stripe/Duolingo-tier taste: "this does not look AI-generated".
+5. SATISFY THE GATES THAT WILL JUDGE THIS PASS. Your work is checked by deterministic gates before it can be accepted, and in measured runs EVERY extra iteration was caused by one of these rejecting the work -- never by the model failing to finish. Clear them now, in this pass:
+   - STATIC ANALYSIS: no syntax errors, no unused/undefined symbols, no lint errors in files you touched.
+   - TEST SUITE: the existing tests must still pass. Run them; do not assume.
+   - MUTATION/MOCK INTEGRITY: tests must assert real behaviour. No tautological assertions (expect(true).toBe(true)), no test that passes whether or not the code works, no mocking the very unit under test, no inline mock data standing in for a real query.
+   - CODE REVIEW: no scope creep beyond the stated task, no dead or commented-out code, no leftover debug output, and follow the conventions already in the surrounding files.
 Deliver the finished, self-verified, genuinely-designed result in THIS pass. Additional iterations should be the exception, not the plan.
 LOKI_FIRSTPASS_EOF